Wednesday, November 27, 2013

Safety Culture - A Little Thanksgiving Goes A Long Way

“Safety culture” is one of those buzz words that gets thrown around in the safety world a lot. The term originated in the aftermath of the Chernobyl nuclear accident to describe those aspects of an organization that are hard to see but affect safety, such as values, social norms, attitudes, and beliefs. The idea is that if you work in an environment where the culture is to be “better, faster, and cheaper” that will effect how you view the risks in your workplace and balance the often competing goals of production and safety.

This seems simple enough to understand, but how do you develop or engineer a safety culture? That’s where the trouble starts. Safety culture is often one of things that we put in the category of “easier said than done.” The reason is that there are numerous influences and components interacting in your organization to create a safety culture. It’s not as easy as just calling a company meeting and saying, “Ok, now we’re going to have a strong safety culture.” That would be like thinking we can fix some of the divisive political issues we face in the US and other countries by simply sitting all the politicians down and saying, “Ok, we need you all to get along and make things better now.” Obviously, it’s just not that simple.

Now, we have no advice for the political arena, but one paradox of building a strong safety culture though is that even though building one is not simple, the things it takes to build the culture, if looked at individually, are quite simple. Take for example recent research into the effect of social exchanges on safety attitudes. Basically, research suggests that in the employer-employee relationship there is a reciprocal relationship where when one party gives to the other, the other responds in kind. So, when you give me a present I feel somewhat obligated to give you something back in return. Conversely, if I feel like you’re being mean to me or treating me unfairly then I don’t feel the need to go out of my way to be nice to you. Furthermore, research suggests that employees typically are not the first to initiate the social exchange – most of the time they wait for the employer to make the first move and then act in kind.

Think about this in terms of a safety culture now – this suggests that you get what you give. If your safety culture is poor there’s a possibility that your employees aren’t seeing your concern for their safety, so they don’t feel the need to care about it themselves.

In the US tomorrow is Thanksgiving Day. It’s a time that we reflect on what we have and we’re thankful for it. What if we also decided to give thanks to our employees for all the ways that they work safely on a daily basis? Certainly no worker is perfect, and everyone does something “unsafe” at some point, but really, at the end of the day, our workers are “safe” more often than not. This doesn’t mean that they are always following the rules, but the goal of safety management is not to get employees to follow rules – it’s to protect people. Further, it’s really those workers at the so-called “sharp end” that create safety by performing their work day-in and day-out in a often varying and many times unpredictable environment. Sometimes this means having to work with (and around) rules that are confusing and many times unworkable.

Now, there are other implications if we did this, but if we thanked our workers for being safe what could that mean for our safety culture? In many organizations workers are seen as the problem, which many workers feel is unfair. According to the social exchange theory discussed above, this makes them less likely to care about safety, making the culture suffer. If we thank them for what they do the workers may begin to see how appreciative you are of their efforts and they are more likely to respond in kind, creating an atmosphere of trust, taking an important step toward building a strong and effective safety culture.


So this Thanksgiving Day take a moment to thank your workers for being safe. How simple is that? 

Tuesday, November 19, 2013

How to Investigate Accidents – Part 2

Last blog we discussed the overall goal accident investigations – which is to improve the system. Too many times investigations stop short of this goal, simply because they found a way to prevent the last accident. This is a big mistake and an important reason why many organizations struggle continually with poor safety performance.

In this blog we’d like to address now how we actually investigate the accident. In terms of the “nuts and bolts” of the investigation there are actually six steps to an accident investigation:
  1. Secure the incident scene
  2. Collect the facts about what happened
  3. Understand the context of what happened
  4. Identify system deficiencies
  5. Recommend improvements
  6. Document the investigation

Lets take a look at each of these steps a little closer.

Secure the incident scene - As you likely have guessed, this is to preserve any evidence that is necessary to understand the accident. Obviously, if the investigation is happening in an organization our ability to secure the scene is limited because the organization likely wants to get things back up and running again as soon as possible. Try to get as much time as you can, but one way to make the process go faster is to take numerous pictures. This allows you to review the pictures later and develop context. Take lots and lots of pictures, much more than you think you’ll need. With today’s technology of digital cameras and camera phones this should not be hard. Remember – you’d rather have to delete a bunch of unnecessary photos than wish you had a taken more photos than you did. And, once you leave the scene, it is difficult to go back and take pictures later. Once you leave, you cannot be sure that the scene has remained secure.

Collect the facts about what happened – This will involve interviewing witnesses, reviewing documents, perhaps review security camera footage, etc. One thing you absolutely need to avoid is to make judgments about what happened at this point. The human mind is really good at making a judgment and then finding information that supports that judgment. In an accident investigation we call that hindsight bias. Unfortunately that leads to a lot of mistakes, especially in the investigation process. The problem, of course, is that no one is 100% unbiased. But this is why we need to actively resist bias, because the hindsight bias has a strong hold on us. Remember – the people involved in the accident did not know that what they were doing would lead to the accident. To them the future was uncertain. Keep that in mind as you collect the facts of the situation.

Understand the context – This is one of the most important steps in understanding what happened and then fixing the system. Behavior doesn’t occur in a vacuum. Things such as organizational goals, the culture of the organization, and even the physical features of our environment have an effect on what we do. Too many times accident investigations focus on human behavior and lead to the conclusion that “you can’t fix stupid.” However, if we look deeper we’ll often find that the environment that a person was in set that person up to make a “mistake” that then led to the accident. If all we do is point the blame at that person we’ll end up dealing with a similar mistake later because we didn’t fix the context, the system that the person was operating in.

Identify system deficiencies – Once you have a good understanding of the facts of the specific incident and the context that the incident happened in you’ll be in a good place to identify what the deficiencies were in the system that allowed the accident to happen. Remember – the purpose of our investigation is to identify problems in the system, not in people. If the only problems we find are with people then we’re missing something.

Recommend improvements – Based on the problems you found, identify corrective actions. For specific items, remember to use the hierarchy of controls that we talked about in a previous blog. For any improvements make sure that you consider unintended consequences. If you implement a new procedure but that procedure actually introduces new risks you might not have improved anything.

Document the investigation – Last, but not least, document the investigation. Usually this is done in a report form for more serious investigations. Each organization will have different procedures and formats that they use. One recommendation though is to figure out what format and forms you want to use now, before the investigation starts. You don’t want to figure out what documentation you’ll need while you’re trying to do an investigation. Consider other stakeholders as well. Does your insurance company need documentation from you? What about OSHA or some other agency? If you’re a contractor, what does your customer want to see? If you work with the public, would there be any information released to the public? Make sure you plan as much as you can now, while you don’t have the stress of an accident on your mind.


As we’ve discussed, investigating accidents is an extremely important part of any safety system. Take some time to digest the information in this blog and the previous one. Share your thoughts in the comments section below. Are there other things that must be considered in the investigation process? The important thing is that you have a clear strategy in place before the accident happens! 

Tuesday, November 12, 2013

How to Investigate Accidents and Incidents – Part 1

Last week, we talked about a common response to failure – blame the people involved. Unfortunately, this all too often leads to an ineffective response. We get rid of the “bad apples” but don’t address the real issues. Then we’re frustrated when we never seem to make any progress in our safety performance. In fact, one of the most effective ways to measure safety culture in an organization is to look at how they respond to failures, accidents, incidents, etc. If all they do is point to how the worker screwed up and needs to be retrained or fired then it’s pretty likely that they have a poor culture.

This begs the question – how do we investigate accidents or incidents? Well, as Stephen Covey said – begin with the end in mind. So, first we must ask ourselves what the goal of an accident investigation is. Take a second and thing about it before moving on – what are you hoping to accomplish by investigating accidents?

Got your answer?

Most likely the answer you gave was something along the lines of – to ensure that it doesn’t happen again. That’s a very important goal and if we have repeated accidents then we should focus in that direction. However, that is not the best goal for an accident investigation for at least two reasons.

First, if you really analyze accidents you start to see a striking truth – most accidents happen because of a unique set of factors that come together at just the right time in just the right way. Certainly there are similarities in some of the factors. If you have two incidents involving electricity you’re likely to find some similar causes. But the way these factors come together is often very unique. This means that if you do what is necessary to prevent the individual accident you’re investigating, but you may not prevent the next accident. No two accidents are the same, so the way you would prevent one accident may not prevent the next one.

This leads to the second point – if all we do is prevent one accident at a time we are being very inefficient with our time? Think about our electrical example from earlier. We investigate and determine that the employees involved didn’t follow Lockout/Tagout procedures. We then discipline the employees and call for additional supervision for these safety critical tasks.

Let’s just assume that these controls are effective enough to ensure that no employees violate any Lockout/Tagout procedures ever again in our organization. Does the employees violating a safety critical procedure like Lockout/Tagout signal perhaps that there might be some troubles in other parts of your safety system? Why did the employees feel like it was okay to violate a procedure? Is this common? Were they ever trained? How are hazards, risks, and safety critical processes identified in the facility?

You see, by answering these questions we start to move beyond simply preventing the accident we’re investigating. We start improving the entire safety system. You see, even though each accident is unique, the causes of those accidents are not unique. Think about it this way, if you have a problem with poor supervision or with an inadequate change management process in your organization, would that be a causal factor in numerous different incidents and accidents? Yes!

The thing is, those organizational weaknesses, or latent conditions, as they are sometimes called, are sometimes hard to see… until the accident happens. That’s why when we are doing investigations we need to set our goals higher than simply trying to ensure that the particular accident doesn’t happen again. Before an accident happens the decisions we make about safety and risk are full of gray areas – we don’t know what’s “safe” and what’s not. But after the accident it’s very easy to see how the decisions we made did or did not allow and the accident to happen. Therefore, we must take advantage of that clarify and do what we need to do to identify and fix any problems we find in the system, even if they were only partially responsible for the accident.

This means that the goal of an accident or incident investigation is to improve your organization’s safety system. If our goal is anything less we may miss key opportunities to reduce risks in our organization, meaning that we are likely going to continue to have problems in the future. Maybe not with the same accident or incident, but with the same underlying latent conditions.

So when you’re doing an investigation set your sights high – we’re out to fix all the problems, not just the obvious ones.

Next week we’ll talk about how we do this. Stay tuned!


Tuesday, November 5, 2013

Don’t Ask Who’s to Blame, Ask What’s to Blame

We recently came across a blog from a local newspaper outlet that we read with interest and not a little dismay. The blog highlights how one of the major regulators in the state, the California Public Utilities Commission, is trying to change its safety culture following a major disaster in the area that it was at least partly responsible for, according to critics. As part of the effort to change its culture the regulator was going to hire consultants to help in the process. The blog is critical of this process, quoting others who point to failed leadership as the root cause, with the fix being “replacing the commission’s brass” as a quick way to change the culture.

Now, we have to say that we’re probably not an unbiased observer on the issues related to hiring consultants, so we won’t speak to that aspect of the blog. However, the implication of much of the blog is that the problems are poor leaders and if we only just replaced those leaders we wouldn’t have any more problems. This line of reasoning is common, flawed, and potentially dangerous.

In response to the killing of three civil rights workers Martin Luther King Jr. cautioned those who wanted justice to not ask who is to blame. Rather, they should ask what is to blame. King knew that if the only thing that resulted from these heinous murders were a few Ku Klux Klan members going to jail there would be no significant long-term benefit. People who were just as bigoted as the others would replace anyone they sent to jail and the problems would perpetuate. The problem isn’t the people, it’s the system that creates the people and allows them to act as they did.

King’s wisdom rings true in so many areas. We often times look for people to blame without considering that it might be the system that needs fixing. The problem is that it makes sense to blame people - it’s easier to just fire someone or tell them to do better next time rather than admit that the system is broken. It’s also a lot scarier to believe that the system is broken. Consider a case of malpractice in medicine – isn’t it scarier to think that the problem is due to a broken healthcare system rather than just a bad doctor?

And in this case, it’s so much more comforting to blame the problem on a few bad apples within the regulatory agency rather than to consider that the people operating in that agency may have to face complex and competing goals every day. The problem of course is that when we just blame the people in the system without trying to fix the system we’re accepting that we’ll run into similar problems in the future because eventually the same pressures that allowed mistakes to be made will push others too close to the edge and we’ll face another disaster with more innocent people suffering.

So, if we truly want justice, accountability, and safety then we need to stop having knee jerk reactions to these kinds of events. We need to ask ourselves not who we need to blame, but what is to blame – what in the interaction between the people in the system and the system itself allowed these mistakes to be made and how can we fix that to ensure that these kinds of disasters never happen again.