An employee gets injured at your jobsite. You make sure the employee gets the appropriate medical treatment, you fill out the forms required by the appropriate regulators and your insurance and then you have an accident investigation.
This is a pretty standard response. The only real question is “how”, as in, how are we going to get all this done?
Some common questions asked:
- How are we going to investigate the accident?
- What methods do we use?
- Who do we get involved?
- How long should it take?
- What do we do with the information we find in the investigation?
There are plenty of resources out there designed to help us answer these questions. There are books, blogs, seminars, training courses and websites devoted to how to investigate an accident. We argue about the best methodologies that will give us the results we want and who should lead the investigation. We discuss what types of events should be investigated – is it just accidents or also other events (e.g., close-calls, near-misses)?
One of the puzzling things to us though is that with all of this discussion about how to investigate accidents there is very little discussion about why we investigate accidents.
After all, accident investigations cost resources. Someone (or a group of people) have to take time out of their normal work schedule, stop producing whatever it is they normally produce, and conduct a proper investigation. Then the organization has to devote resources to fix whatever problems the investigators find.
So, investigations come with a cost. Are they worth it? Why do an investigation?
Now this probably seems like a trivial question to a few of you and we totally understand why – everyone knows that the reason to investigate accidents is to figure out what went wrong so you can prevent it from happening again.
Let’s look at this a bit closer.
There’s a key assumption underlying this reason – accidents are caused by negatives that must be found and removed from the system (or prevented from having their effect). Something broke, someone failed in some way, etc. Therefore, the job of an accident investigation is to find the thing that broke (i.e., the root cause) and fix it (or replace it). Once fixed, the system is safe, and the accident can’t happen again.
There are some problems with this line of thinking though. The idea that bad things are caused by something failing or breaking is something that works great for machines. If you’re trying to figure out why your clock stopped working you look for the part or parts that failed and fix/replace them. But, if we’re dealing with people and organizations, it doesn’t work that way. People do things that help them achieve success (or avoid failure).
Therefore, if someone does something that we say leads to an accident, to call that a failure or an error is often a very narrow and limited view.
To illustrate, let’s use an example –
someone worked on a piece of machinery without properly de-energizing it (i.e., locking it out). From the perspective of a safety professional that is a failure, because they failed to follow a rule. But the person didn’t do what they did because they wanted to fail. They acted in a way that they believed would help them achieve success in the environment they were in. Therefore, from another perspective, there was no failure.
Now, this doesn’t excuse the behaviour, but look at how the reason we investigate (the why) leads us to specific conclusions.
If we take the perspective of finding the broken part, we would often stop here. The employee broke a rule, that’s the broken part. We might decide to go a little further, look at the supervisors, the managers, etc. But our inquiry is about why the person didn’t care enough or didn’t know enough to do what was necessary to stop the accident.
Therefore, we always end up in the same place – blame. Someone, somewhere failed. People are the enemy of safety. Corrective action – name, blame, shame and retrain.
What happens if we take a different perspective –
the reason we investigate accidents is to learn about how our system is working and improve it? This is based on the idea discussed above, that people do things that help them achieve success, but sometimes these things also cause accidents. Success and failure (accidents) have the same causes. Therefore, looking for what broke and trying to fix it doesn’t work. Instead, figure out how things normally work, i.e., why things are normally successful, to understand why it didn’t work this time.
Notice right away that by changing why we investigate accidents it changes how we investigate accidents.
For example, if we accept the idea that success and failure have the same cause then methods designed to identify the faults or failures that led to the accident won’t be very effective. We need methods that help us understand success, not things that only model failure.
Looking at our example above, if our goal is to learn and improve it leads us down multiple paths. We can start asking questions about equipment, working conditions, incentives, communication, culture and how all of these elements worked together to lead to the behaviour in question. In the process we are free to ask questions unrelated to the accident. We aren’t constrained by only looking for “cause” anymore. We look at the instance as a special case of normal work, so if we find something we don’t like, we fix it, just like we would during non-accident conditions.