Predicting Safety Outcomes: Learnings from Big Data with Dr. Chuck Pettinger
LISTEN TO THE EPISODE:
ABOUT THE EPISODE
An excellent interview with Dr. Chuck Pettinger exploring what Big Data can teach us about safety and preventing the next potential injury. Exploring what information and insights are available now and predictions about the future of safety in the digital space.
To learn more about SIF Prevention: https://www.propulo.com/blog/tag/sifs/
READ THIS EPISODE
Real leaders leave a legacy. They capture the hearts and minds of their teams; their origin story puts the safety and well-being of their people first. Great companies ubiquitously have secret productive operations. For those companies, safety is an investment, not a cost for the C-Suite. It’s a real topic of daily focus. This is The Safety Guru with your host Eric Michrowski, a globally recognized Ops and The Safety Guru public speaker and author. Are you ready to leave a safety legacy? Your legacy success story begins now.
Hi and welcome to The Safety Guru. I’m very excited today to have Chuck Pettinger with us, who’s coming to us from a company called Predictive Solutions, to share some really cool topics and focus areas around the future of safety, really in terms of how we can predict and prevent injuries using data and information. So, before we get into the meat of the topic, Chuck, can you share maybe a little bit about how you got into safety?
Yeah, it’s kind of a long, roundabout road, but I actually started when I was in graduate school and I was doing research with a professor. His name is Scott Geller. Yeah. And we had several different grants that we were working on. And then I got assigned to a Nesh Grant, which is the National Institute for Occupational Safety and Health. It’s OSHA’s research arm. And I worked on that grant for a couple of years. It was entitled The Critical Success Factors for Behavior-Based Safety.
Then I was on an MSRA grant, which is the mining side of OSHA. So, I really got indoctrinated into the safety side of it. Now, funny enough, I got my degree in psychology and you might think, well, how would that relate? But, you know, safety is all about people, right? And psychology is the study of people. So really, my focus is more on the industrial-organizational side of the psychology equation. So how people interact, how leaders can motivate the right types of behaviors and those types of things.
So, you know, once I finish my graduate program, I started doing consulting work and. Right. You know. Thirty years later, I’m still getting out there and helping people save lives, so that’s kind of my beginnings.
That’s an awesome story and great focus and great mission in terms of what you’re doing. So maybe if you can talk a little bit about some of the data that you’ve seen over the years, you’ve attracted a lot of different organizations. You’ve had a chance to analyze a lot of data and observations. Are there any major learnings that you’d like to share with safety leaders that can improve their effectiveness?
Oh, yeah. I mean, there’s a ton of information out there. And for the last 15 years, we’ve been collecting data and we have half a billion safety inspections, audits and observations in our data set. And we have data scientist’s feedback. You know, several years ago we partnered with Carnegie Mellon, which is in Pittsburgh, and they actually had the Language Technology Institute. And these guys are the ones that built the program to beat Kasparov and chess.
I don’t know if you remember Big Blue, you can Google it, but it was the first time a chess computer actually beat, you know, a grandmaster. And since that time, they have gone on to help IBM create Watson, which is the supercomputer, but. What they did is they took all of our data and we said, well, can we predict injuries based on leading and lagging indicators? And at that time, we had enough data that they could run some programs and they came up with really interesting things.
And some of them, I had a hunch. Right. But I didn’t really have the proof. But I actually now can say definitively that there’s four safety truths. And kind of the first one is if you increase the number of inspections, audits or observations, you’re going to decrease your injuries, which makes good sense.
Yeah, of course. That’s what everyone assumes. If we get out there and we look at more stuff, we’re going to be able to be more proactive and get ahead of the curve and be able to impact that. Now, the problem, no matter how big your organization is, you don’t have enough incidents to be able to have enough statistical power to say, if I do X, I’m going to get Y. But that’s why we have this big data.
So, we can say that increased inspections, decrease injuries. Safety, truth number two. Now, it’s not all about just increasing the number of inspections. What we found is that you need to have a good diversity of observers out there looking at things. So interestingly, what we found is that the locations that were part of the study that actually only had the inspections being done by safety professionals actually had a higher injury rate.
Conceptually, it makes sense to me, because it would mean to me that people in operations maybe haven’t really bought into what’s happening. It’s only in safety.
Excellent. Yes, exactly. That was my thought. Everyone’s like, well, you know what? I don’t really buy into that because they’re safety professionals. They should know what they look at. But if the safety people are doing it who own safety, right.
Nobody. Yeah, because it really needs to live in the business.
Exactly. So, the more diverse your observer pool, the lower injury rate.
So, if I may. Is there anything when you talk about diversity, is it also in terms of different levels of seniority within the business, different type of functions? Have you seen that or is it more in the dimension of safety versus operation?
Yeah, you know, we didn’t get into that level, but we are able to identify who the safety professionals were within the locations that we did the study. You know, I would love to seen I mean, I would hypothesize the higher up we get to be engaged, the lower the incident rate is. But I mean that’s another future food for thought. So, I really like yeah, I really like that. That part of it. So, safety truth one increase inspections, audits, observations.
Yeah. You’ll decrease injuries. Number two, have a lot of different people out there looking, not just your safety. Number three, and I affectionately call this the pencil whipper safety truth. If you start seeing 100 percent safe observations, we call those unicorn observations, you’re going to have more injuries. Right. Which makes sense. Yeah. So, what that’s telling me is not that you have bad employees, but you have a bad broken process because no one’s actually using that data.
And honestly, from the employer’s perspective, if nobody is looking at this data, how much time and energy am I going to put into doing it? So, they’re just checking off boxes to get this one more task done. Makes sense. Yeah. That number for the fourth safety truth is kind of the exact opposite. If you start seeing an increase and the number of risky things that you see, you’re going to have more injuries. Now, this is to me, even worse, because the data is staring us in the face and not acting on that data.
So again, those things. I mean, if you were to ask me what are some truths about going out, doing inspections and audits, I would probably come up with those. But we actually have the data to say definitively, you need to get people out there looking, not just safety people. You’ve got to get a diverse group of people out there and you have to act on the data. And that is what’s going to make a difference in health and safety of our employees, and I love how you’ve been able to demonstrate it through data, and that’s obviously that the value of having a repository that captures so much information from so many different organizations.
So, I think this is really interesting. Intuitively, like you said, it makes sense. But now you can prove with data what organizations really need to do.
Yeah, it’s fascinating.
And can you share maybe some thoughts around the importance of leadership engagement? So, you’ve alluded a little bit around in terms of diversity, but can you double click on that one a little bit more?
Certainly. I really think it flows right into our last discussion and that if. The leaders are not acting on this data. It creates what I call a venomous cycle. So, what happens is people get all excited, they roll out a new process or they have a new checklist and go out and they do observations and then they never hear anything about it. Right. And then they go out again. They go through the motions. Nobody says anything.
And then pretty soon I’m just checking off boxes. Mm-hmm. I was at a conference in Indianapolis and I was talking about this. And a guy in the back room says, yeah, I have this whole inspection thing worked out. I take home a handful of these checklists. I give them to my daughter. She fills them up for me to bring them in. And I’m done for the whole year.
Right. So, I don’t blame that guy again. It’s all about the process. So, if people are not acting on that data now, let’s reverse that. If I turn on something and my supervisor looks at it, acts on it, makes a change within our systems or processes or environment or training. And I see that I have an immediate value proposition. What was that, a generation facility and every Monday morning, the senior-level leaders and safety folks and the safety group, they look at their data.
And what was interesting is the supervisor was walking around and he was just, you know, doing his daily walk. And a guy walks up to him and says, hey, come over here and observe me doing this, OK? Right. And he says, watch me walk up the stairs. And, you know, those metal gratings on those. Right. Right. This had come up. So, it’s like this toe grabber type of thing.
So he goes, watch me trip up the stairs. I want you to record that. So, you know that following Monday, the leaders looked at that, said, hey, how long’s it been like that? I don’t know. Right. Maintenance guys are in the room, said, hey, you know, Joe, can you go get that taken care of? Sure. We got some more grading in the, you know, the trailer. Let’s go knock that out.
So, what they did is they actually took a before and after picture and posted that in the break room. So, and they titled it Your Observations in action. I love it. So, again, if the employees see that and tie what they’re doing out in the field to things getting done, you get better quality data. The conversation is much better. And people don’t think that this is something that, you know, you’re trying to hunt down people who are doing things wrong and shame and blame them.
It’s all about trying to find gaps in our processes and systems and then close those gaps. And so, we create what I call a virtuous cycle. Right. I stole that from economics, but still works here. Right. You see the value in it. You do more of it. And that’s a simple way of improving your process. And all it takes is leadership to act on that data.
And so often that’s exactly what I’ve seen, is organizations are not closing the loop. People don’t realize the value in it. It’s not being reinforced and they just see it as time and effort that goes absolutely nowhere.
So, yeah, and, you know, I really think a lot of companies may not be at that level yet, you know what I mean? Because I think some companies are still at the level of complying with regulations. Right. So, I’m supposed to go out and observe my employees. That box is checked. No one ever said I have to actually follow up on that.
So have meaningful conversations, do anything about it. Right.
They don’t consider that impact on the safety culture. And, you know, as long as I’ve been doing this, those organizations who are proactive in their measures versus, you know, rallying the troops after somebody gets hurt. Right. Those proactive organizations are much safer. They have a much more open. They have a much more even compliant organization because they see the value in it. Right.
I agree it makes sense. So, you have a chance to work with a lot of cool technology, understand really what data can bring to the table. If you had a crystal ball and you had a chance to look at what the future of safety will bring us with all the advances in technology. What are some of your thoughts around what might happen, what might be available for us in the near future?
Oh, my goodness. I mean, all you have to do is go to any of the large safety conferences and you will see the Internet of Things. If you have not heard that you’ve been living under a rock, you can Google that phrase. But there is so much smart technology being woven into safety equipment, smart technology and vests that can detect your body core temperature, whether you have fallen, the ambient temperature, your location. So, I think if I were to guess, we’re going to have a lot of these things giving us data in real-time now, we actually have to be able to use that data.
It’s absolutely useless if nobody uses it. You know, I have this Weight Watchers app on my phone. I’m going to delete it because it’s not working.
Right. It’s the app that’s not working.
Yeah, exactly. It’s the apps not working. Right. So, you actually have to act on that data. So, here’s what I see. You wake up in the morning and you speak. Alexa, what’s the forecast and then you say, Alexa, what’s my safety forecast right now, Selcuk? Well, today on your route to work, you know there is a car crash, so we’re going to reroute you so you can have a safe and speedy journey to your job.
And then when you get to your job, the work that you’re going to be doing requires you to work outside and it’s going to be high humidity and really hot out today. So, we’re going to send you some reminders. Now I get to the job and I’m looking at repairing a particular generator, and I’ve never seen this before, so I have my Google glasses on and I tap them and I call up my supervisor, say, hey, Josh, you know what?
I’m looking at this. Have you ever seen this before? Oh, yeah. Yeah. Chuck, you know, you got to do is go on the other side, open up this and you will see it. And he can write in my field of vision and circle things and say, oh, hey, watch out right here, because this can really capture. Right. And then this data can be transmitted immediately. So, if I am getting overheated, he might say, call me up, send me a text, Chuck, take a break, go grab some water.
So, all of this data is going to be happening in real-time. And that is kind of what we’re bracing ourselves for. We have smart homes, we have smart cars, we have smart TVs, and that’s to help them humans, I guess. I don’t know.
But to make that work, you need some form of aggregator. You can’t have ten thousand different apps that see if different infrastructure information to different systems. It somehow needs to roll into one view. I would think so that you can do something that meaningful with that data. I’ve seen too many organizations collect tons, tons of data, but have no meaningful way of using that data to drive decisions and actions.
Oh, my goodness. Yeah, the data is going to be streaming. Right. And I’ve seen these vendors. There’s this one vendor that had a belt that you were at that can detect if you’re lifting safely, if you have fallen and it can stream the data live to you. But look at that and take that live. Nobody. Yeah. So, you need some sort of algorithm. You need some sort of data scientist to go in and tell you.
All right. So overall of your employees, we’re starting to see a trend in this area going give safe lifting training. Right. And that is the key. What we need to do as a community, as a world is develop open-source APIs so that we can pull in each other’s data into different pieces. You know, almost every organization uses the Microsoft package, empower by is one of them, and they’re making it easy to visualize data.
It’s a drag and drop. It’s not liked the old, you know, databases that we used when we were kids to short our CD collection. That’s when we actually purchased music and not just streamed it, but so we do need to learn from this big data. And that’s where the data scientists can help us.
So, I think that that leads me to the next set of questions. And your organization is really at the leading edge of predicting using information for safety outcomes. And I’d like you to share a little bit about what that looks like, but I’m really interested in that, really in this space around. How can you predict safety outcomes and how have you seen that meaningfully help leaders keep their workplaces safer?
Yeah, it’s I mean, it is an ongoing art and a little bit of science. I mean, you can do any sort of prediction, but it doesn’t necessarily mean it’s predicting the right thing or predicting it specifically enough. There is a lot of incident or incidents where we’ve actually predicted down to the body part, the type of injury and the how serious it is. Right. But it is too specific and they focus only on that.
And then when it can happen, you know what you don’t want in the first place, they’re not quite sure if they did it or the model was wrong. So really, what I see big data doing and these learning algorithms is telling you, all right, so you have this large organization, it may be global, it may just be one facility. But within your realm of responsibility, here are some things that look different than everybody else in the world or some things that look like other organizations who have also had injuries.
So might check that out. It’s like having the, you know, the light on your car. Something’s going on. But you need someone knowledgeable to look at the data and decipher it. And that’s where we get the engagement from the employees, from the leadership team to look at that and say, hey, what’s going on? I remember there is an organization I worked with and we would redly flag locations that looked like they were going to have an injury now.
Every January, all of the locations would be red-flagged. And you know, all the managers, guy, this doesn’t work, this is like voodoo magic, and what we learn from that is, well, in December, everyone takes off for vacation. The number of observations go down, the number of findings go down, the number of closed issues go down. So, to the computer, something is completely different from November to December. So, in January, you’ve got to look out now.
Now we have to teach the algorithm that in December and maybe in the summer, you’re going to have a decrease in activity because of vacations and people taking time off. So, it isn’t the, you know, magic eight ball, right. It still needs the human to look at that and make sense of the data, because without that, it’s going to it’s not going to be as impactful. It still needs us to say, oh, yeah, that makes sense.
Now I know why.
But it gives you a really helpful clue in terms of where should I be focusing my efforts? Yes, drive something that could potentially happen. And rather than complain to say, gee, that incident happened, you should be happy that maybe you change something in the context that prevented that from occurring.
Yeah, it’s kind of like going fishing, go to a giant lake and just drop your line anywhere and try to catch a fish. What these predictive models do is say, I’ve found you the best fishing spot you need to go here and you’re more likely to catch fish. So, for us, that’s kind of what we have to do is say, right, why is this data different now? And that’s where we get the employees coming in saying, oh, I know why we really started getting a nurse on-site and now people are turning in more first aid cases and going to get band-aids.
So, our near misses and our first dates have skyrocketed, not because there’s more, it’s just because it’s easier. So that type of thing helps us understand the system and the processes that are putting people in a place to either do things that are risky or to do things that are safe.
Interesting, and what are some of the limitations that you’ve seen on the data that’s available today and this type of context, what are some of the things that leaders need to be aware of?
Yeah, and, you know, there is no silver bullet right now. You can say that you want this predictive model, but really the limitations sometimes are you don’t have enough data. The data is not accurate. And a lot of times I’m talking to different companies and I get to know what they’re doing. Like Chuck, we want to do these predictive analytics. And I just tell them, I said, you’re not ready. Then you’re going to get is not accurate enough.
You don’t have the culture to be able to use that information and you’re going to be wasting your money. So really, again, what we have to focus on is what is going to produce the best outcome, which is people going home without getting hurt for sure. That’s identifying those serious incident and fatality precursors. So, start identifying those things that could potentially give you a lifelong impairment or end your life. Focus on those, identify the things that are putting people in risky situations.
We can be a lot more proactive and predictive. It’s not just given me a paradox chart of, you know, worst to best. It’s actually helping you define where you need to focus on. And sometimes the limitation is the data just doesn’t make sense. I mean, there’s plenty of times and working with our data scientists and like Chuck, these three variables came out tied together, you know, maybe something like, you know, day of the week, eye color and favorite ice cream.
And, you know, and you’re going to have absolutely no idea how these ties together. Right. And it just may be some random variables. You know, again, to me, the data is secondary to the conversations we can have that the data helps us focus in on. So, I’m all about data and all of our research. But even more importantly, if we can translate that data into things that are meaningful to our employees, have conversations with them, we increase that engagement, we increase the visibility, we increase the transparency.
And really what we’re doing is creating a more open and factfinding culture. I don’t know if that makes sense, but it makes a lot of sense. I think it’s really well set because at the end of the day, it’s how the leader takes that and has the right conversations. This is just an extra trigger to me at least, that that helps prevent a manager from taking place once you’ve got the basics, right.
Yeah, precisely. Precisely.
So obviously, to be used, the data needs to be as close as possible to real-time. Can you share how important that is and any other learnings that you’ve had in terms of data and in the safety space?
Yeah, I mean the more real-time we can get at the better it is. Sometimes it’s very difficult because sometimes we just only calculate monthly and so we roll it by month. But that means we’re only looking at things 12 times a year. So, if we’re only looking at those 12 times a year, that limits our impact. So, this as real-time as possible. I was doing some work with a company that makes diesel engines and they have locations all over the world.
And these two guys were at a location in Hazard, Kentucky. Funny. But so, these guys, one of the younger guys was setting up a hoist to lift up one of these big engines. Old timer walks by and says, hey, you know, Josh, I know that that is too heavy for this hoist. You know, it’s a fourth down engine. That’s only it always is only rated on two tons. So, the young guy, you know, Josh looks at the box and says, oh, now it says right here in the crate, this is only two tons.
And the other guys go. Now, I know for a fact that is a four-ton engine. So, they look up the specs and sure enough, the engine is four tons and the engine was mislabeled on the crate. So, they record that data and they have what they call a severity alert. So went into to the system as a high alert. The corporate safety guy over in Indianapolis got this information and he got on the phone, started calling the distributors.
And sure enough, there was a batch of engines that went up that were mislabeled and by that afternoon the distributor had called up to different locations and had already started sending out new labels. So, from that morning to that afternoon, they identified something that could potentially been a serious incident or property damage or ruined, you know, a brand-new engine. Yeah. And, you know, when people see that happen, that excites them. Right. I just not only saved my potential coworker, but maybe I’ve saved the whole company a lot of money.
So, you know, that just shows you the power that getting this data in real-time can be. And that’s what is exciting for the future of safety.
Absolutely. And it also goes back to what you were talking about before in terms of people seeing the value of the observations of what you’re what you’re seeing, what you’re putting in as information. If it’s acted upon, the feedback loop comes back. This is really the epitome of a learning organization in my mind.
Yes, definitely phenomenal. Chuck, it’s been fantastic having you on the show. I think you have some really cool ideas. If somebody wanted to reach out to you to get more thoughts, more background in terms of how to take their safety to the next level, really by using data, what would be the best way for them to do that?
Well, you can email me Chuck Pettinger, at Predictive Solutions dot com. You can look up our company, Predictive Solutions, dot com. And I often talk at AFSC and ASSP and all the regional safety conferences and then KSE up in Canada. I spoke there several times, so I’m usually around and you can find me on LinkedIn as well. So, I really appreciate the opportunity to share some geeky data stuff.
But you made it in a fun, interesting way and in very meaningful ways. I appreciate it.
Thank you. Oh, no worries. No worries. It’s fun.
Thank you for listening to The Safety Guru on C-Suite radio. Leave a legacy, distinguish yourself from the pack, grow your success, capture the hearts and minds of your team’s—fuel your future. Come back in two weeks for the next episode or listen to our sister show with the Ops guru Eric Michrowski.
The Safety Guru with Eric Michrowski
More Episodes: https://thesafetyculture.guru/
C-Suite Radio: https://c-suitenetwork.com/radio/shows/the-safety-guru/
Powered By Propulo Consulting: https://propulo.com/
Eric Michrowski: https://ericmichrowski.com
ABOUT THE GUEST
Chuck has over 30 years of experience designing, implementing and evaluating culture step-change initiatives. His major interests include developing large-scale corporate behavior change initiatives, assessing industrial safety cultures, using advanced predictive analytics to develop leading indicators and conducting organizational Leadership Workshops.
For more information: www.predictivesolutions.com