Welcome back to the Safety of Work podcast. Today, we discuss whether “pointing and calling” improves reliability of action.
As our workplaces become more automated, it becomes the task of human workers to monitor the automated actions. At times, this may require a physical response or action on behalf of the human worker. So, while the physical load of workers has been lessened, their mental and emotional load has increased.
Tune in to hear us define pointing and calling and the ensuing discussion about its efficacy within the workplace.
Topics:
Quotes:
“You point your index finger directly at that thing and you say aloud what that thing is currently showing”
“But this pointing gesture also acts as a cue to trigger this attentional shift towards the information.”
“The researchers did not state clearly what their hypotheses were. For those of you out there who are doing research, this is a big no-no when you’re doing an experiment…”
Resources:
Drew: You're listening to The Safety of Work podcast, episode 73. Today, we are asking the question, does pointing and calling improve the reliability of action? Let's get started.
Hey, everybody. My name is Drew Rae. I'm here with David Provan. We're here from the Safety Science Innovation Lab at Griffith University. This is The Safety of Work podcast. In each episode, we try to ask an important question about the safety of work or the work of safety. We have a bit of a look at the evidence surrounding it.
David, what type of work of safety are we looking at today?
David: Drew, we're looking at a process today called pointing and calling. We'll explain that in a little while. But just for a bit of context, in our increasingly automated workplaces, the role of human operators in many control room or controlling style situations involve monitoring the status of certain processes and making certain responses in relation to information that you’ve presented. At times, this can require just a simple physical response such as pressing a button or touching a screen in response to a particular situation.
We all have different ideas about what the reliability is of humans performing these types of actions. While the physical load for operators has been minimized over time because of all of the different technologies, touch screens, buttons, and control panels, operators still have to manage an enormous amount of information that's fed to them through lots of different information systems.
By improving the person-computer interface, we now can contribute to reducing the amount of information processing load on the operators. But at the end of the day, the person operating the system will always still be responsible for making critical decisions and carrying out critical local actions. If the person makes an error in this type of control operations and if the system is unable to detect and correct this error, then it can have catastrophic consequences.
Drew, therefore, in my opinion, there is a lot of effort in safety science that continues to explore ways to try to reduce this risk of human error. People can't see on the podcast, Drew, but I'm holding up my hands in big inverted commas when I use the label human error.
Drew: I think that's one of the more difficult conversations to have in safety. A lot of the stuff we do is about preventing human error. A lot of the resources that we started using are saying that human error doesn't really exist. That's something that makes it really hard to have clear and frank conversations about how much we do rely on humans to get things right.
We've got humans as part of the system. The performance of those humans matters but there are some types of performance that we really don't want. We've had a big backlash against calling it human error because it's really a system error where humans are part of that system. We still want to have that human part of the system operating as well as we can. Often, we want to do that through system features like the design of the interface, but also how we train people to behave.
David: Absolutely true. I refer to it when I think of the role of people in these company systems in that people are incredibly high-performance parts of the system. People can deal with a whole range of different information and different novel types of environments. We want an infinite amount of algorithms to make sense of the world around us, but we're not necessarily equipped with machine-like reliability for our actions. This idea of 100% reliability of human action doesn't exist and we know that. We know that when we talk about making error-tolerant systems. Much of the effort has gone into this interface design between the operator and the information communication technology or ICT.
We've been looking at this for decades and decades since we've had this different interface and different display units of how dud it gets represented and how that stimulus is responded to by the human operator. But like you said, it's really important to think about the role that people play because we're people after all. We've got to process that information and take action.
Drew, in this episode, we're going to talk about one such operator process. It's called finger-pointing and calling. The article we're going to talk about talked about FPC as this idea of pointing and calling. It's an operational procedure that's implemented in safety-critical industrial settings as far as I know almost exclusively in the Japanese industry. It’s designed to prevent human error. It's referred to as shisa kosho or shisa kanko in Japanese which is this point and call.
Drew, do you want to tell us a little bit about what is pointing and calling?
Drew: Sure. David, you and I might have run into shisa kanko in the same way which is riding around trains in Japan. I certainly encourage any of our listeners, if you're in Japan and you're on a train, try to get out the front and look through the glass door or window at the operator, the driver at the front. You can see this in motion. The driver is going along and each time they see a signal, even if it's just a green light, they point a finger directly at it and they say the Japanese word for green, the Japanese word for red, or the speed limit that they've gone past.
The idea is you're looking out for things that you're supposed to see whether it's a sign, a meter, a button, or a light—things that you need to check or control. You point your index finger directly at that thing and you say out loud what that thing is currently showing. If it's a speed meter, you point at the speed meter and you say, 60 kilometers an hour. If it's a pressure gauge, you point at the pressure gauge and you say 30 psi. If it's a red light, you point at it and you say red. If you're using it to change control, then you point at the control and you say what the control is. Brake handle—point at the brake handle before pulling the brake handle.
We'll talk a little bit about why that's considered to be a good thing but David, do you want to tell the story about when you've seen it?
David: Yeah. The paper went on to say that this idea of physically checking, pointing, or touching what you're doing and saying out loud what the status is, what you're about to do, or what action you're about to take in relation to the system is done in healthcare settings and surgery. It's done in pre-flight processes in aviation. But it's designed as a two-person checking activity where you say it out loud and the other person confirms out loud.
In this situation, we're talking about just a solo operator, an individual cognitive process, or a physical cognitive process that the person goes through. I've also seen it on trains. When you mentioned the idea for this episode, I recalled that there are some really good YouTube clips on this. Many people probably won't be unless you're an Olympian, you probably won't be traveling to Japan this year. For anyone who wants to have a look at what this looks like, you can use YouTube pointing and calling, and you'll be able to get some good little clips of what it looks like.
Drew, you could also try this for yourself. After you listen to this episode, you could also try this for yourself safely I suppose. Maybe this is safer. Next time you're in your car, you can drive along and every time you see a speed sign, take one hand off the wheel, point at the sign, call out loud 60 kilometers an hour, point at the gauge, 60 kilometers an hour, and confirm the action that you've taken—stop sign, brake, red light, brake. What you're doing is you're painting at the stimulus that you're responding to, you're saying out loud what the status of that stimulus is, and then you're also saying out loud the action that you're taking.
Drew: Actually, in Queensland, at least one of the first recommended lessons when you're teaching your kids to drive. They say you take them out on a drive and point out each of the hazards, or get them to drive along, calling out each of the hazards that they can see. Not quite the same because the last thing you want is someone actually pointing out all of those things while learning to drive.
The pointing is actually considered to be a key part of this task. But as you said, David, it's used extensively in Japan. As far as I know, it's been for at least since the early 1950s, but there's very little written about it. Obviously, a lot of what's written about it is in Japanese, but we've got access to the abstract. We can see that even in Japan, there's not a lot of academic research. All of the stuff that's been written starts out by saying that it's well-known that finger-pointing and calling is effective and therefore, we're doing this study to explore why it works, which is interesting.
If you've got something that's been around in Japan for so long that people are so confident it works, but it hasn't spread elsewhere in the world—we might get into that a little bit later in the discussion because if the reason it hasn't spread is lack of evidence, then that doesn't explain how lots of other safety practices have spread. But it does seem to be popular and accepted as a safety thing in Japan.
David: Yeah, absolutely, Drew. I think you meant the 1950s. There are a few papers referenced in the paper that we're just about to introduce. There seems to be some research done in the mid-1960s. It seems to be the most heavily cited initial research around this pointing-and-calling activity. But then there was also some work done in the '90s where participants were required to respond to one of five visual stimuli. This was performed in random order, five different stimuli with five different responses presented to participants in random order.
These participants were asked the way they carry out this finger-point and call. Point and call out which of these five random stimuli it was and then take some action, or do nothing and just make the response to the stimulus. Basically, this study in the mid-90s showed that doing the finger-pointing and calling, significantly reduced the errors of taking the wrong action in response to those five random stimuli.
I think that this evidence has passed but some of these laboratory-type studies—which we're also going to talk about one of these types of studies today—seem to show that the process at least doesn't make things less safe.
Drew: At the very least, we have quite a plausible theory for why it should work. You got a number of different things that are going on. What I'm doing here is just repeating a claim that's been made in these papers. Take it with a little grain of salt as a plausible theory rather than something that's been proven.
The idea is that the first thing we have is confirmed visual contact at the target. We're deliberately looking at something within bringing that thing into our memory. The way that I refer to it is there is a continuous orientation of attention. The idea is we're not just seeing it, but we're also saying it and that makes it more salient in our minds, easy to be remembered. We're doing more than just seeing. We're seeing multiple senses. We're looking while speaking. We're presumably hearing ourselves. We're moving our arms and doing a muscular movement.
Pointing at things is a human gesture that really increases the activation of our mind towards the thing that we're pointing at. All of these steps make it really plausible that we are avoiding that common problem of seeing something without noticing it, seeing it without fully processing it, or seeing it, and then it immediately slips our mind. We're taking a very deliberate act to make sure that it is salient in our memory and that we've looked at it properly.
David: The theory or the claims in this research talk about it being a very deliberate action sequence. We're almost slowing down the cognitive processing that we're doing because we're having to take the time to point, to speak, to hear ourselves, and then to take action. It almost forces us to slow down and this action sequence can increase—as the claim is—that accuracy gets promoted because of this deliberate sequence.
The previous research on this topic is early literature. It says that most of it concludes that pointing and calling may be useful for improving the reliability in situations that emphasize accuracy and speed. That's generally the claim made in the literature, although we're going to see some things around the speed that are maybe counter-intuitive as well.
The literature also says that this pointing and calling might also be particularly useful for some habitual behaviors. If you think around the house which is I don’t know how people do but sometimes you go, keys, wallet, phone, sunglasses, leave the house. You put your hand in your pocket when you confirm that your keys are there and your hand on your head for your sunglasses. But this idea of things like have I locked all of the doors in the house? Have I shut all the windows?
Some of these things that we just do as part of our everyday tasks, forcing ourselves to point and call out, could potentially pick up errors in some of these habitual behaviors. If you thought about before I left the house, I'm going to walk around the house, I'm going to point at every door lock, I'm going to point at every window, and I'm going to call out whether it's open, closed, locked, or unlocked and check it, it's logical to say how that could prevent oversight of some kind.
Drew: You're saying increased accuracy, low speed, drastically increased self-conscious silliness before you leave the house.
David: I guess so.
Drew: It's worth pointing out that this theory says that the opposite also should be true. If what's really more important is speed than accuracy, then we don't want to do finger-pointing and calling. Obviously, if you're doing something like playing a first-person shooting video game where whoever twitches first wins, the last thing you want to do is stop and point out, hey, that's a bad guy before you shoot. More seriously, if you've got situations where there's a risk of error from operator overload, some of these papers suggest that you deliberately slowing down also ends up increasing the workload if there are too many things to point at, too many things to process.
David: Drew, maybe we'd talk briefly—I don't think I have too much data—but we've talked about the logic of the action sequence and the deliberate nature of that. Let's talk really briefly about what might be going on or what the literature suggests might be going on inside our brains in these situations.
There's one model of human behavior that's referred to in the paper, which I just realized we still haven't introduced yet, but we're almost there. It suggests that our actions are controlled by two cognitive systems of your life. One is this is a contention scheduling system and one's supervisory attentional system. Don't worry too much about those. Basically, they're saying that this finger-pointing gesture has an effect on the perceptual stage of information processing. As you said, Drew, it allows us to notice what we're seeing—not just seeing something but actually perceiving and noticing what it actually means for us.
By pointing the finger, this visual attention gets orientated and focused on this target stimulus and we are more certain about the perception of the information that's being displayed. It's the detection of the visual target. This pointing gesture also acts as a cue to trigger this attentional shift towards the information. At any point in time, we're noticing a whole lot of stuff in our environment. If you think about when you're driving along, you’re noticing that the cars around you, all the different signs, the buildings, the for-sale signs on the house, the person walking their dog. But all of a sudden, if you point at a particular piece of information within that broad environment, it causes this attentional shift directly towards that piece of information. Physical gesture creates this really targeted attention to that information more so than if you didn't actually do the pointing.
Drew: David, let's talk a little bit about the paper. I should say upfront that I wanted to do this topic. I pulled out the most recent and best paper talking about it. This is not the most thorough study. What I'd like to highlight from that is the opportunity we have here for more research. I know some of our listeners are at universities or are doing research. This is one of those things where it would be very easy for someone to set up and do the very best study ever on finger-pointing and calling without too much effort. Just have a look at this study. Know that it got published in a good journal and know that maybe you could set up something that is a little bit stronger.
The paper is called The Effects of Finger-Pointing and Calling on Cognitive Control Processes in the Task-Switching Paradigm. It was published in the International Journal of Industrial Ergonomics in 2013. I have to admit, I haven't looked up the authors in detail. Did you, David?
David: No, I didn't. I just looked at where they were from.
Drew: The authors are from Osaka University and from the Institute of Nuclear Safety Systems—Kazumitsu Shinohara, Hiroshi Naito, Yuko Matsui, and Masaru Hikonob. The paper goes in saying that the effectiveness of pointing and calling has been widely recognized and it is there to prevent human error, but it's not well-understood what the cognitive processes are. This paper is setting up to tease apart some of the steps, some of the reasons why it's effective.
David, can we just dive in to describe the method?
David: Yeah, let's go. Let's talk about what the research did.
Drew: Okay. Each participant basically has a left button and a right button. What's going to happen is there are going to be two numbers on the screen in front of them. Each number is just a single digit. There are three different ways in which these numbers can be different from each other.
One of them is that one can be a higher number than the other, the other way is one might be higher on the screen, and the other way is that one of them might be bigger. In each variation of this study, there's going to be a rule that tells them which number is the one that they should pick. Maybe the rule this time is to click on the higher number. Maybe it's to click on the larger number. Maybe it's to click on the physically large number.
The participants first get shown the rule, they click ready, they get shown the two numbers, and then they click either left or right. We’re measuring the preparation time—the time before they say they're ready. We're measuring the time—the time to the first click. We're measuring the accuracy—whether they clicked it correctly. Is that a fair summary, David? Anything I have missed there?
David: No. Other than there were four different conditions that you left through. I'll just briefly mention those. That's basically what they did. They got given a task rule, they said when they were ready to go, and then they clicked their response. They're either doing a single task which means they did, say, 30 consecutive trials just with the same task rules. Pick the number that's higher on the screen and do it 30 times with the same rule, or the mixed rule situation where it was a randomized one of the three. Pick the number that's higher on the screen. Then, it might be to pick the number that's in a larger font size or pick the number that's numerically higher. It’s like that.
They are the two different conditions. For each of those, they either did it with the finger-pointing and calling or they did it without the finger-pointing and calling. Everyone did the same, went through the same exercise in four different conditions.
Drew: When we talk about finger-pointing and calling, it’s what they were doing during this rule-stage of the task. The rule would flash up, like pick the largest digit. They would point to that rule and say the rule out loud. That was the pointing and calling part of it.
David: Perfect, Drew. They also had a survey that they did straight after the end of each of the trials. It was 26 questions and a 7-point scale which was a scale that was about subjective mental workload assessment—how hard was the task of that before me basically. The covered areas are mental demand, physical demand, time pressure, performance, effort, frustration. They did this survey at the end of each of the trials for the different conditions because the researchers also want to understand this claim in the literature about whether finger-pointing and calling make tasks more mentally demanding, or higher workload because of the extra activity associated with the task.
Drew: The researchers did not state clearly what their hypotheses were. For those of you out there who are doing research, this is a big no-no when you do an experiment. You've got to say in advance what the primary claim is, but I think we can infer from the introduction in the paper and the way they wrote it.
Their first prediction was that the finger-pointing and calling were going to make things more accurate and that it was going to be particularly helpful when we went into this mixed task where the rules kept changing. Then, we had the second remeasure of time which will help us explain why the finger-pointing and calling is more accurate.
What in fact they found was that the participants were all very good at the tasks. Over 20 participants, each doing 30 at a time, and then 4 sessions of that 30 so a total in the end, of 8000 times that they did the task. Only 2.5% of those were incorrect—a very, very low error rate. That low error rate didn't really seem to have anything to do with either the task or the finger-pointing and calling.
Sorry, I misspoke this slightly. The error rate was higher in the mixed rule, but it wasn't different between the finger-pointing and calling and the not-finger-pointing and calling. The mixed rule task was harder, but even with the harder task, the finger-pointing didn't really help with the accuracy.
David: Drew, you think some of those numbers like the single rule was a much lower error to the multi-rule, which means that if someone is doing something repetitively 30 times and the rule is the same, then keeping that in the working memory and getting the right response seemed more reliable than when the rule kept changing every time a new task flashed up, which makes a little bit of sense, but the error rate was three times higher in that situation even for a simple task like this. When you think of going from picking the number that's this or picking the number that's that when you keep changing up even what seems like a simple task, you're introducing three times the error rate.
Drew: The first basic idea that point-and-call which is going to improve the accuracy doesn't appear to be true in this particular study. I don't even remember if it was shown from the previous study that this study cites. What they did find were some interesting things with the workload and with the timing because it certainly wasn't the same that pointing and calling had no effect at all.
It was obviously changing the way people were going about and thinking about the task. This was most noticeable in the mixed rule situation. Sorry, I'm misspeaking here, David. I might just get you to take over for a moment.
David: No, that's fine. In the mixed rule situation, this is how I interpreted the results. If people remember the method, basically, a rule flashes up on the screen and the participant had to say when they were ready to be shown the numbers so that they could actually perform the task. It flashes up on the screen, chooses a higher number, and they say, yeah, got that, go. They either pointed and called out what the rule was, hit the button, and said I'm ready to go, or they just processed it without pointing and calling and said, yeah, I'm ready to go.
It seems from the result that with the finger-pointing and calling, people were ready quicker. It was almost like they had read in between the lines. My conclusion from the results was that they had a way to approach the processing of the task. It was as soon as that rule flashed up on the screen, they pointed, they read it out, they said I'm ready to go, and off they went. Whereas when it just flashed up on the screen, they maybe couldn't focus their attention as quickly or they couldn't process the information as quickly without that pattern of behavior of pointing and calling, but it was definitely intentional preparation time which they described in the paper. If people have the time to prepare, to be ready to go for a task, the pointing and calling seemed to speed up that preparation time.
Drew: With a simple task, pointing and calling is helping them be faster at being ready to click on the right thing. When it came to the more complex task, the people who were pointing and calling were spending longer on that preparation screen. But they were then faster once that screen was gone to get the right result.
It was like it was leading them through the process of loading the rule into their mind. They were a little bit slower about loading the rule, but once the rule is there, they're faster to activate the rule when the next screen comes up. Whereas the people who didn't have that assistance through the point and call were faster in saying that they were ready. But then they had to stop and think about what was the right thing to click on.
David: Drew, anything else in the findings that you thought was interesting? We’re now at the point where we go, okay, we've got this practice of finger-pointing and calling used extensively in high-risk settings in Japan. We've got a bunch of literature that simply talks about accuracy. This study didn't find an accuracy finding but found this processing speed finding around pointing and calling. Anything else in the findings here that you think is worth talking about?
Drew: I guess the main big takeaway from this paper is the whole big claim for pointing and calling is that it's a way of reducing human error. This is widely accepted throughout the industry in Japan. It's widely practiced. Even regulators accept it and so presumably, check up on it. Then, we have these studies that are really fairly mixed, and the most recent most rigorously conducted study just shows no difference in accuracy.
It probably shouldn't surprise listeners of the podcast that the authors come away concluding that the study provides evidence that finger-pointing and calling is effective and therefore should be encouraged to be introduced to real-world working situations as a tool to prevent operational errors. Obviously, the research does not show that. What it does show though, is that there is definitely a cognitive difference in people doing finger-pointing and calling to people who aren't.
It's not a practice that has zero effect. It's just not proven to be a useful effect. That's the type of result that definitely should get you thinking, hey, there's something going on here. Maybe it's helpful. Maybe there are different situations where it isn't helpful. Certainly, this is worth doing some more experiments to find out—is it generally helpful and what tasks it helps with.
David: I think so, Drew. It's definitely a little bit fascinating because I must admit before I read this paper, I thought, oh, when you look at the information that is generally available outside of the academic literature around finger-pointing and calling seems to talk about that it is the single biggest cause of the high reliability of the fast rail industry in Japan. It all boils down to this one practice of finger-pointing and calling for not crashing trains.
Then, when you dive into the literature, it's obviously not that compelling but there was definitely something going on in this laboratory setting. In many of the things that we talked about some time in the podcast, maybe there's nothing going on at all. But here is something. I think it would be possible that the tasks in the laboratory are just a simple rule on a screen and pressing the left and right buttons are such a simplistic action and stimulus that the accuracy is always going to be highly accurate anyway once you put humans into that task.
I'd love to see a study in a real-world environment with some more complex stimulus and some more complex human behaviors like driving a train, flying a plane, controlling a hazardous system, and trying to actually understand if it works because there's something really going on here.
Drew: I don't know about complex real-world environments, David. I'm thinking that if you are a PhD student and looking at a little bit of experimental research and safety, there are lots of practices that we really have gotten a good reason to suspect that they would work at all. Here, we have something where a good experiment might actually find something. Why don't you fire up a copy of Microsoft Train Simulator, set a few signals to read, get someone to sit in front of a train simulator screen, count error rates, and see how it works?
David: I don't know about train simulators, maybe Grand Theft Auto, or something like that. I think a gaming environment could be interesting. I’ll happily supervise that study and I know Drew would use any excuse to play more video games.
Drew. Absolutely. Are there some practical takeaways?
David: If in your workplace, you've got a situation where an operator has to check information on a visual display unit, they have to memorize that content, and they have to perform one of a range of different actions in response to what's displayed on that display, then pointing and calling may change the accuracy or the speed at which that they actually process that stimulus and make that action. Like we've said, we can't be probably more conclusive than that, but it would be an interesting thing to explore within your workplace.
Drew: The second one that is worth thinking of is pointing and calling as an individual mindfulness exercise. We've got lots of things that we ask employees to do that are supposed to get them to recognize hazards at the moment. The worst example I'm thinking of is someone who sits down and fills out a risk assessment card just before swinging a ladder into a powerline. My immediate thought for that accident was what if instead of having to fill out the card, they just had to look up and point up, look down and point down, look in all four compass directions for hazards, and point at the hazards? Maybe that would have been more effective.
If you're dismissing this as a practice that is it just seemed silly to make workers do this, then think about the other things that we make workers do that they probably think are just as silly and ask yourself why something that has been widely adopted in a country that has a very good safety reputation hasn't spread. What is our reluctance? How do we choose our safety practices?
David: Yeah. Maybe we just don't want to look silly. Maybe do it as part of work but I think that's a fantastic practical takeaway. I think that the possible idea that having a group of people stand around at a worksite, write a risk assessment form, or complete a risk assessment form versus having a group of people standing around at a worksite and point and call out hazards may create a very different attention and management of those works site hazards. That's another potential research project for someone as well.
Drew: If the problem is just looking silly, David, I just want to remind you that our last episode was about high-vis and it included a discussion about the importance of wearing pink or possibly watermelon.
David: Absolutely right, Drew. I'm wearing pink high-vis shirts when I'm walking around sites. The very active pointing and calling, waving your hands around are also going to make you more visible as well. It's going to make sure people know that you're human because you're moving your arms around. Not only will you process the information that you're doing by pointing and calling, but you might also avoid getting hit by a car in the process.
Drew: Before we wrap up, I do want to acknowledge from the last episode. We had this whole episode about high-vis and we never once mentioned colorblindness. I've had a bit of listener feedback saying that that is actually an important consideration of contrast. Contrast still works for people who are color blind. That's something we're going to need to have a look at and maybe cover in a future episode. I just want to do acknowledge now from the feedback that that is something important that we actually missed from the episode.
David: Good point, Drew. I absolutely hadn't considered that. I suppose the diversity of individual needs in our workplaces is something that obviously should factor into our safety considerations. Absolutely. Apologies to those listeners for missing that aspect and thanks for the feedback.
Drew: Invitation to listeners. Do we have any Japanese listeners you know from the podcast, David?
David: Sure we do. I know we got downloads in more than 100 countries now, Drew, so I'm expecting Japan to be on that list.
Drew: This is not a wide invitation, but I'd really love to hear from any of our listeners who have direct experience with point and call who might be a bit more familiar with the evidence, for and against, to be able to read those Japanese papers other than just the abstract like we had access to.
David: Great, Drew. The question that we ask today, in this episode was, does pointing and calling improve the reliability of action? Drew, do you want to have a go and answer that?
Drew: Specifically for the paper we reviewed this week, no, it did not. But the balance of evidence, there is something going on there that we need to look into further.
David: Great. That's it for this week. We hope you found this episode thought-provoking and ultimately useful in shaping the safety of work in your own organization.
Please join us in the discussion on LinkedIn, drop us an idea for a future episode into the portal, or send any comments, questions, or just any general feedback to us directly at feedback@safetyofwork.com.