Tuesday, March 10, 2015

Accident Investigations Are Error Provocative Environments


Editors – To hear more about our thoughts on accident investigations, join us as one of our team members presents Beyond Root Cause – A New View of Accident Investigations at the Orange County ASSE Professional Development Conference on March 18, 2015. This week’s blog is a preview of that presentation.

A staple of the safety profession is accident investigation. When things go wrong we marshal up the resources to figure out what happened with the hope to prevent it from happening again. Take for example a relatively recent injury at a client’s site, where an employee used a 21” radial saw and had his finger amputated. Obviously this is an outcome no one wants, so the organization organized an investigation, led by the safety professionals. In this case the safety professionals found that the saw had a sign indicating that the saw was “out of order”, but the employee used it anyway. While using it, the employee shut the power down on the saw and reached toward the blade to make an adjustment. Unfortunately the saw blade was still moving, amputating one finger and lacerating another. The accident investigation report counseled the organization to provide more training to employees in the shop, and to ensure that no employees work in the shop alone. They also recommended replacing the saw with a “safer” alternative.

One of the pioneers of the human factors field, Alphonse Chapanis, was one of the first to identify that sometimes people make mistakes not because of deficiencies in the person, but because of the environment they are working in. For example, one of his early achievements was to reduce aircraft mishaps in World War II by simply changing the design of toggle switches to make them easier to identify what the switch was for. Simply adding a small rubber tire to the toggle switch for landing gear reduced incidents of accidentally putting up the landing gear to almost nothing. Chapanis, and other human factors pioneers, identified early on that many times errors can be simply chalked up to a poorly designed environment – i.e. an environment that is designed so that the human must overcome their humanity in order to succeed. These environments are sometimes called “error provocative”, because they tend to provoke errors in people operating in that environment and simply changing the environment can go a long way to reducing mistakes.

We submit for your consideration that accident investigations are error provocative environments. We think that asking someone to investigate an accident, particularly in the way accidents are commonly investigated, is putting them into a situation where they are highly likely to make a mistake. Consider what we are asking a person to be able to do in an accident investigation:

  • Be objective. The idea that people can be completely objective at nearly any point in their life is simply unsupported by any empirical evidence that we are aware of. All of us take all of our experiences, education, emotions, personality, culture, etc. into every situation we go into in life. This all colors our perceptions of the world we live in. Even in a highly sterilized environment, such as scientific research, Kuhn found evidence that our worldviews (Kuhn called them “paradigms” to represent patterns of thought in a field) drastically affect what gets researched, what questions get asked, and therefore, what answers are deemed acceptable by a scientific community. Note that all of this happened without the conscious awareness of the researchers. If this is true in scientific research, where emotion should play not part, how can we expect objectivity in the highly emotional environment of an accident investigation? Add to this that often we are investigating environments where we had significant influence before the accident. As Manuele noted, often this puts us into situations where the investigator has a vested interest in downplaying their role. In some cases, Manuele points out, the investigator may be writing their own performance appraisal with the investigation. How can we expect objectivity in such an environment?
  • Explain the complex in a simple way that facilitates action. The point of an accident investigation in the minds of most safety professionals is to lead to action, so we can prevent the bad thing from happening again. So the organization expects a simple explanation of what happened that facilitates getting things fixed. The problem is that accidents are often quite complex. Many aspects of the normal environment came together in an unexpected way. Sure we can point to many things and say “if only this didn’t happen then the accident wouldn’t have happened”, but where do we draw the line. Often we end up with arbitrary stop rules on our investigations where we just decide where to stop the investigation. Unfortunately this often leads investigators to solving the problems they wanted to solve before the accident. So the investigation is led by the need to solve problems, not by the need to learn.
  • Be thorough and efficient. In all environments we have competing goals, but accident investigations are hyper-examples of this. We must be thorough and really dive into what happened, but we also have rules that require us to turn in investigation findings within a certain time frame, so we must be efficient. Unfortunately you can’t be both completely thorough and completely efficient. Something has to give.


If we go back to the employee amputation we can see examples of these problems. Upon our analysis, the accident investigation missed a lot of really important lessons. The saw was purchased years ago and was left in the shop, partially set up. The employees in the shop didn’t want to use the saw because most were intimidated by the size of it. So the saw was not really out of order, it just wasn’t completely set up, and all of the employees knew this. One employee felt comfortable using the saw. Guess which one…

Now we do not blame anyone in the organization for any of this, whether for the accident or for the investigation. The problem is that, in both cases, people were put into situations that invited mistakes. Unfortunately, the problems that led to the accident, although complex, are easier to solve than the problems that led to an accident investigation that missed some key learning points. As a profession, safety professionals need to rethink the way we investigate accidents. If we just “go with our gut” and investigate without really thinking through and challenging ourselves, we will often come up with a poor product. We need to develop environments that facilitate learning. Some quick examples to consider to make your investigations better are:

  • Get others involved in the process whenever possible. Diversity in the process often leads to better outcomes. This includes getting the employees who were involved in the accident involved in the investigation, if possible.
  • Use investigations as learning processes. Searching for “the” cause may actually make the end product of the investigation worse, because it might lead us to stop the learning process. Investigations should be about learning, and often after an accident the organization is willing to learn, so take advantage of them.
  • Start thinking in terms of systems. The old ways of understanding organizations just can’t capture the complexity of the organization. There’s a ton of great resources out there, such as from Meadows, Checkland, Leveson, and Senge. Find something that works for you and start the learning point today!

No comments:

Post a Comment