Dominic Furness, a consultant at Human Reliability, presents an overview of Human Factors Critical Task Reviews (also called safety‑critical task reviews or analyses) and examines why these workshops sometimes succeed and sometimes stall. He outlines the regulatory backdrop in the UK (HSE roadmap, COMAH requirements) and the core methodology – hierarchical task analysis, failure analysis, and identification of performance‑influencing factors, followed by recommendations via the hierarchy of control and organisational learning. Furness introduces a lightweight application of the Functional Resonance Analysis Method (FRAM) to explore socio‑technical variability, illustrating how functions (e.g., passing, holding the ball) and their variability affect overall system performance, and maps these concepts onto the task‑review process. He derives practical success strategies: ensuring frontline staff availability, building rapport and a safe space, allocating sufficient time, and leveraging in‑house software that visualises hierarchical task structures to boost engagement. A client case study on a hose‑pressure‑testing task demonstrates how pressure, shortcuts, and inadequate preparation can hide risks, yet a visual hierarchy can surface hidden failures. The talk concludes that safety‑critical task reviews are fundamentally socio‑technical activities, and their quality can be studied and improved through FRAM‑informed reflection and better integration into organisational processes such as procedures, job aids, and training.
Chapters
Thank you. I'll try and make up some time so we're not too late for lunch. But yeah, so I'm Dominic Furness. I'm a consultant working at Human Reliability with Mark at the front there. We've come together, and also you might notice David Embury. Many of you might have heard of him. You'll see from the methods that we will discuss that there's a Sherpa influence sort of coming through, and also some of the software that we use in industry to help us sort of deliver those methods. So I want to focus on the human factors critical task reviews, otherwise known as safety critical task reviews or safety critical task analyses. These are really important in the UK, part of the regulatory system, which I'll describe. But what I want to do is start to think about the performance variability of applying critical task reviews. So when we go out and we engage with different companies, sometimes it feels like they go really well and we get lots out of them. Other times there's challenges and they might not go so well. Sometimes we're asked to come in and do training and mentor and support other people working through their processes and their sort of different degrees of success as well. So using FRAM to try and explore some of that area and thinking about safety critical task review, not just in technical terms, but socio-technical terms, because it's run by people, we rely on software and artefacts to deliver it, and all of that goes into some of the makeup into whether it succeeds or it doesn't. So, just a little bit about the company. So, we mainly do consultancy services, really looking at human factors in procedural tasks. So, that applies to many different industries. But predominantly due to the regulations that we work in, we engage with oil, chemical, and gas energy type companies. We also have a background in, at university, so we like to keep up research. We've currently got some projects going on in healthcare looking at artificial intelligence and managing the human performance aspects of that. Education and training. And also development of software to help people apply some of these techniques. This is the human factors roadmap from the Health and Safety Executive in the UK. I won't go into the details here because we're a bit short of time, but just to note that safety critical task reviews is a requirement by COMA sites, so any sites that are going to under risk of major accident hazards that have multiple fatalities or big environmental damage will need human factors input and comply with these sorts of processes. So how do we achieve that as a company? Well, in the middle here, we have the human factors critical task review, which is the main substance of what we do. So that generally composes of a hierarchical task analysis, then a failure analysis, and then looking at performance influencing factors that either increase or decrease the likelihood of these failures happening. Now, as we go through that process, we'll uncover sort of different risks and make recommendations via the hierarchy of control, you know, try and design them out if we can, or if we can't do that, then we might have to look at management of performance influencing factors. But sometimes not noted explicitly is this idea of organisational learning we have there. So when we're holding these workshops with operators, management, process safety experts on the site, giving them sort of an environment where they can talk together and discuss and share problems, issues of non-compliance, whether the procedures are working right for them or not. So that, that's a really important aspect of this which sometimes doesn't go noticed, especially after we've done the risk analyses and, and task review. We can output that into procedures, look at job aids, and also training aids as well for people as part of managing the human performance aspects of that. So generally, a workflow would be selecting a task, then we get the procedures from the company, and we do a preliminary task analysis, and then we can go in and sit down with the operators and go through, make sure they're happy with— that's actually the way they do the work, so the work is done. Sometimes the procedures are based on work as imagined, and we want to bring those two together. Actually go through a walkthrough, so they point out things on site, which is also important because sometimes they can't always articulate what the problems are, and us as human factors experts, we can identify things maybe they can't. and then deliver the report and then have some sort of integration. So we don't want this just to be a detached safety thing that sits on someone's shelf. We want to try and integrate it into the normal organisational processes. So part of this paper that I'm presenting, just to go over the method a bit. So I observed colleagues doing safety critical task reviews. Reviewed previous reports that we had done. We had discussions and also carried out some task reviews as well as part of the data collection. And then thinking about some of the variability in the output of these and how sometimes they go, you know, well and sometimes not so well, we decided to use FRAM just to explore that. A bit, but in a light-touch way. It's not like a full application of the FRAM method, which can be quite complicated. And I've published before on how to use FRAM for looking at positive things, so how to maximize the quality of processes rather than just looking at it as Some people use FRAM and they only look at trying to prevent negatives happening. I think that's quite in keeping with the Safety 2 ethos. So just to briefly outline how we might think about this positive resonance. So the first thing we do with the FRAM approach is we might identify the functions of the system. And then we think about their variability individually, so how is each function performing? And then how do these functions perform together? So here, I mean, it's only a toy example, if you like, but it's based on a football team performing well or not so well. And so rather than outlining the individual players, having a goalkeeper, defenders, and things, We can outline some of the functions, like they execute the passes well, they hold onto the ball, they keep cool under pressure. All of these different functions, if they're performing well, then the team we would expect might perform very well. If some of them are not performing well, then we might have problems, the team might lose. And then we have maybe critical connections between these. So if there's, if for example, the team we have has someone that's really good at crossing the ball, and then we have strikers that are good at doing heading, then we have like a critical sort of path there where they, where we can win the game by adopting that strategy. So this is about how teams and socio-technical systems can perform well. We can think about the— this is just a visual illustration of how we can think about some of the variability of these different functions and how they come together and affect overall system performance. So, this is the FRAM analysis that we came up with basically. We populated it with lots of functions which are involved in safety critical task reviews. That one in the centre contains a lot of the functions within a workshop and walkthrough, but it's influenced by things upstream and it affects things downstream, affecting the quality of the, of the report and also of the procedures, training aids, etc. You have to look at the paper to get a bit more of an idea there, but what we try to do is try and extract out from that some strategies for success. So making sure that frontline staff are available so we can talk to them, otherwise we're not going to even engage with that process bridging work as imagined and work as done. And this can happen sometimes. It might not be that the frontline staff are completely unavailable, but we might only have them for an hour in a 2-day workshop, and it's just not enough to engage with them. Building rapport and creating a safe space so people are willing to share issues of non-compliance. You know, this is quite sort of like a— fairly soft factor, if you like, but it can have significant impact on our assessment of the risks involved in some of these tasks and potential recommendations. There's some other ones there as well which I'll skip over, but again, if you look at the paper, they're all listed. I just wanted to tell tell you about a story we had. So we, um, we were called into a client and we needed to— they already had safety critical task review processes that they were happy with at one level, but they wanted further training into sort of like potentially how they could improve their performance in these. And ahead of time, we did a preliminary analysis of one of their tasks When we got there, we found out from the people that actually attended that they'd already gone through this task and they wanted to do something else, but we'd already done preliminary work, so we thought we'd try and go through it just quickly, but we were a bit apprehensive about that because we thought, if they've done it already, then we're probably not going to find anything. So, we presented it to them, just the hierarchical task analysis. And all of a sudden, they started discussing things and issues with it, details about how to organize their work, maybe reducing— it was a hose pressure testing task, and they had two different types, so potential failures that you might mix them up when you're calibrating them. And the guy in charge said, "Yeah, really organizationally, we want to reduce that to just one." so they didn't have the threat of mixing them up. And, yeah, and all of this was coming out of something that they'd already done. So we were thinking, you know, there must be something going on here. They've not performed as well as they could have done. But they— I think they themselves were under quite a lot of pressure. They had hundreds of these things to do. They didn't get a lot of resource and time to do them in. They sort of like created slight workarounds and shortcuts, so they didn't do a separate task analysis, they just imported their procedures into an Excel table, and they didn't recognise sort of different activity types and then follow on the different failures associated with them. Now, some of the nice things about the software that we use that's been developed in-house is that all of this is sort of built into the logic. So, when— and I mean, from experience, I think if you put up a visual hierarchical display, it really engages with people, so people start pointing at boxes and say, "Oh, can you move that to the left?" and "That should go in the middle." I just don't think you'd get the same sort of interaction and feedback if it's just in a tabular form. Then we've also got the straight link to the tabular form, so we're not— it's not any sort of extra work for us, and the activity types and the failure modes that are linked to them are sort of all integrated. So it makes it a lot better. So that's some of the technology that impacts staff engagement within these processes. But also the group dynamic as well. So because they were under pressure, they were looking to force through the processes quickly. So I think trying to be more efficient than being thorough, if we're looking at that sort of trade-off. So that's just in conclusion. I wanted to get across the point that socio-technical safety critical task reviews shouldn't be just seen as a technical activity, but the quality of their deployment is a socio-technical one, and we can study that and we can try and reflect on the quality and improve those processes. Thank you.
