Accident Investigation: Back to Reality

While safety is important, putting safety first sends the wrong message to our organization. Here's why.

Could the incorrect identification of causes lead your investigators into the "twilight zone" of accident investigation?

You safety guys are always slowing down production and driving up costs by demanding those pointless accident reports for every little thing that happens. When are you guys going to get real so that we can get back to work?

Sound familiar? If so, comments like these from your line organization may be more than just whiny excuses or good-natured banter. They might be indications of something gone wrong with your accident investigation program.

Obviously, not investigating accidents would be a major blunder to any safety program, were this element not included as one of the most critical for success. On the other hand, I believe that in some instances, the accident investigation process can lose sight of reality and become a detriment to a company"s safety improvement program.

Reasons to Investigate

Should we investigate all accidents? Contrary to how it may sound in the paragraphs above, I believe the answer to this question is a resounding "yes." I"ll even go a step farther: I believe that all hazards, even concerns, should be investigated to determine all the causes for their existence. Why? To take corrective action in order to prevent an accident from happening or happening again. Accidents are caused by unsafe acts and conditions. Accidents are prevented by eliminating or controlling unsafe acts and/or conditions. Knowing even one act or condition that hasn"t been eliminated or controlled is being aware of an accident waiting to happen. Although there is always a priority list with our work loads, not making a conscious effort to do our best at the above would be negligent.

There are two other reasons for investigating all accidents. The first is to ensure that injured personnel receive the workers" compensation benefits that they are entitled to when injured on the job. Like any insurance program, it is there for use when needed, and it is a no-fault system. A poorly written report with improperly identified causes can result in an injured employee"s claim being denied.

The other reason to investigate all accidents is to protect your company (and yourself) from false claims. Whenever an accident is reported, it must be investigated and a report written, even if the outcome is that nothing occurred. Your investigation report stating this conclusion can become a very important document. Yes, all accidents must be investigated. There can be no other answer.

If all accidents are being investigated, what"s the problem? I believe it is the identification of causes that leads some investigators off the path of reality and into the "twilight zone" of accident investigation.


I am often amazed that, even among safety professionals, there are different definitions for what should be basic terminology for the profession. Let"s make sure that we are all speaking the same language.

Accident - There are probably as many definitions for this term as there are companies. The dictionary defines it as "an unexpected and undesirable event." I like to say that an accident is something that happened that you didn"t want to happen and you don"t want to happen again. Many words say the same thing, but my definition does emphasize the need for prevention (thus, why we investigate).

You will notice that neither definition says anything about losses. An accident that did not result in a loss is still an accident, and could have resulted in a loss, aka "near miss." I hate the term "near miss" because some people equate it with something insignificant when, in reality, it may very well be the opposite. Even if a particular near miss could never result in a major loss, the same causes, if allowed to go uncontrolled, could result in other, much more significant problems.

Immediate Causes - These are the unsafe acts and conditions that resulted in or could have resulted in an accident. They explain why an accident happened. The thought processes of some people lead them to insist that something like "cut with a knife" is an immediate cause. This statement explains what happened. It does not explain why it happened. Examples of immediate causes for being cut with a knife would include using the wrong tool for the job (unsafe act), failure to wear cut-resistant gloves (unsafe act), or dull knife (an unsafe condition).

This last cause brings up another point: If something is intended to be that way, it is not a cause. Thus, a sharp knife can"t be the cause; a knife is expected to be sharp. If the employee in this example is supposed to use a knife for this job, then "wrong tool for the job" is not the immediate cause, either.

Root Causes - Root causes are the management system causes that result in unsafe acts being committed or unsafe conditions existing. These are causes that affect your whole system; thus, their correction not only affects the single accident being investigated, but many other future accidents, as well. For that matter, they could also be the root causes for other management and operational problems.

The root cause is the most fundamental cause that can be reasonably corrected to prevent recurrence of the error. This last point is most important. For example, it is not enough to say that the root cause is that supervision was not at the scene at the time of the accident. It is not reasonable for supervision to be everywhere at one time. If the supervisor was not required to be at the scene before the accident happened, then his or her absence can"t be a cause for the accident.

One more assertion about causes: You have to be able to assign corrective action to them. "Wind" is not an immediate cause because the wind is supposed to blow, and "act of God" is not a root cause because you can"t assign corrective action. Wind damage is caused by inadequate storage, inadequate design and the like. The management system causes that allowed these and other immediate causes to exist are the root causes.

Problem Areas

Inadequate definition for the terms noted above can certainly cause problems for your accident investigation program. However, I think there are other areas that result in even bigger stumbling blocks. One of these problem areas is the amount of time spent in investigating an accident. Granted, sometimes possible liability issuesmake us feel it is necessary to take extra time to ensure that all ducks are where they should be. However, too often, we waste time looking for something that isn"t there. This is especially true when a team is assigned to do the investigation. Of course, there are situations where the team approach is best, even required, but, for most accidents, an investigation team is not necessary.

Example: There is snow and ice on the sidewalk. Someone slips and falls. It is natural for snow to occur; therefore, we know that snow cannot be the cause. Thus, our focus should be on determining if our snow/ice removal efforts are adequate. Was snow/ice removed and ice melt applied? If not, why? Was slip-resistant footwear in use? If not, why? Should it be required? For most companies, these questions could be answered in 15 to 20 minutes. It does not require sending samples of the snow to a lab to determine if it is, indeed, slippery, nor does it take a team arguing on whether or not to paint the sidewalk another color so that the ice might become more visible.

By the way, I like to use the above example to show how the degree of injury is usually a matter of luck. The snow/slip situation can result in no injury or a fatality or anything in-between. The injury depends on how the person lands (or bounces). To be proactive, the causes must be identified and corrected when the hazard is first realized, not after a serious injury occurs.

The time required to perform a thorough investigation is the amount necessary to have the "whys" answered. When there are no more real answers, the questions should stop.

Another problem is investigators trying to find the root cause when there is none. How many times have you heard someone say that there is always a root cause? This simply is not true. Let"s say I were to drop my pen on the floor. If I am not paying attention when I reach down to pick it up, I can hit my head on something.

Some people will agonize for the longest time trying to find the root cause for my inattention. In reality, there is probably no management system cause for it. There could be for example, supervision having just chewed me out in front of my peers (a very bad management technique).

However, in many cases where inattention is the immediate cause, there is no root cause. It is not correct to say that "distraction" is the root cause. That would be synonymous with "inattention", so we would be saying the same thing. The root cause is never the same as the immediate cause. Remember, the root cause is the management system problem that allowed the unsafe act to happen or to exist.

Let me give you an example of how an accident investigation can go wrong. Description of accident: An employee was helping at a line blender station and was asked to obtain a large plastic bag. He went down the stairs and over to the bag rack on the south wall. It was empty. He turned around and went north towards the bag rack on the north wall. As he passed the large mix bowl, he slipped on some water on the floor. The fall resulted in a cut above the eye and a fractured wrist.

This accident was investigated by a team of five individuals who met for four to five hours (20 to 25 man-hours) and discussed such things as the urgency created by a problem with the blender, the hard cement floor, the requirement to mop up water immediately, putting dye in the water to make it more visible and the employees" fat feet (EEE width) decreasing the coefficient of friction. Talk about being in the "twilight zone"! What do you do to correct the last item. Take a knife and trim his feet down to a D width?

The team did end up identifying the leaking mix bowl as the root cause. In reality, this was the unsafe condition; however, in this case, it was also known prior to the accident. Therefore, the correct immediate cause is use of inherently hazardous equipment, meaning that the safety hazard with the equipment was known and its use condoned by supervision prior to the accident.

The root cause deals with why supervision did not fix the leaking bowls prior to use. Because the team did not ask the right question, this root cause was not identified and corrected. In other words, although the leaking mixing bowls were repaired following the accident, the reason supervision did not take the appropriate action prior to the accident remains a mystery and could very well result in other known hazards not being corrected until other accidents occur.

Cause v. Blame

Corrective action must always be appropriate for the causes identified. This is easier said than done for a lot of people. There seems to be two types of supervisors: those who blame the employee for everything and those who think that their people never do anything wrong.

The words "cause" and "blame" are not synonymous. A person can commit an unsafe act that causes an accident but that does not mean that the accident was his fault. There could very well be a root cause that forced the person to commit the unsafe act. There is also a difference between making a mistake and a flagrant violation of company safety policies and procedures. "To err is human," but if someone cannot get into trouble for purposely violating safety rules, then your safety program likely will. Correcting these very different causes and the varying degrees of severity associated with them is not an easy job; however, being fair and consistent can alleviate much of the difficulty.


I am sure you have your own examples of runaway investigations. Although all accidents must be investigated, almost all can be completed in a relatively short period of time, simply by asking "Why?" at the appropriate times. Time consumed trying to find hidden causes seldom adds anything of value.

Likewise, identifying causes that are incorrect is a waste of time, no matter how quickly the investigation was completed. I would not expect all supervisors to have the expertise necessary to identify all causes correctly. However, I would expect them to call on their safety professional for assistance in this endeavor. All safety professionals should be, or should be well on their way to becoming, proficient in accident investigation and cause identification. OH

Larry W. Sorrell, CSP, is a senior safety engineer/scientist with the Propulsion Group Division of Thiokol Corp., Brigham City, Utah. He has more than 14-years" experience with this aerospace firm, with accident analysis and trending being one of his primary responsibilities. He holds B.S. degrees in biology and in occupational safety and health from Utah State University.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.