Focus on critical risks, you say. That’s all good but what does it mean? And what does it look like from a governance perspective? And why do it at all?

The why
There are several reasons why this focus matters.
Boards get swamped with information and, due to the breadth of issues they are dealing with, that information is necessarily summarised. So, it’s crucial that they can be sure they’re focusing on what matters most. Time is precious. Don’t waste it talking about the fact that we’ve had three people hit their thumb with a hammer, when there is a near death experience somewhere else in the business to be concerned about.
We’re also trying to avoid the tyranny of the frequent. It is easy to be seduced by those events that happen often. But these, almost by definition, are not going to be significant. This is made worse if we report at a senior level on every individual incident that happens, because availability bias makes people overestimate the real frequency and importance of those things they see reported a lot (see Governance Series – Meaningful Reporting).
There are many examples of major events that were preceded by long periods of small stuff going well – Deepwater Horizon being the most obvious. If you focus on minor things, you will get better at those things, but to the detriment of the big stuff. And that’s what we really need to avoid. Safety losses are not like financial losses – there is no straight-line equivalence scale. Ten losses of $100k equals $1m, but ten cut fingers does not equal an amputation. We have to look at the high potential risks.
If we focus on the risk, rather than its outcome, we stand a better chance of identifying problems in our approach to those major events that could lead to catastrophe. They simply don’t happen often and so are harder to spot. We have to be more careful about where we look.
The what
What is a critical risk?
There are a few definitions floating around, but little that formally defines what a critical risk is.
The one I see most often, and what I usually recommend, is risk from those hazards that have the (realistic) potential to result in a fatality. This might change depending on your risk profile and the type of work you do, but the key thing is to get a clear dividing line that excludes stuff that’s trivial. And although you can have major disabling injuries like amputations that are not fatal, but certainly not trivial, the chances are that the sources of these will be picked up in the fatality definition.
Having defined and identified them, what’s next? I see many organisations stop at this point and do little more than awareness raising. What you are looking for next is assurance programmes to make sure that those critical risks are being well-managed. But, to do this, there is another level of criticality prioritisation to do. This is about understanding what controls you focus on. Every significant risk will (hopefully) have a lot of controls in place. If we try to put robust assurance around all of them, we run the potential of clogging up systems through sheer volume. So, we need to prioritise again.
This is a fairly subjective process in most cases. The key question to ask is, “Which of these controls contribute the most to mitigation of the risk?” For each of the identified critical controls, develop a management and assurance plan that gives you confidence it is effective.
The how
How do you go about a management assurance plan and what do you share with your Board?
Your management and assurance plan has two components – making sure the control is good enough to do the job it’s supposed to and then confirming that it is working. Sometimes these are referred to as control design and control operation.
When thinking about design, you should clearly define:
- what the control is protecting against
- what the safety function is that it must perform and
- to what standard should it perform it.
These are generally easier to define for engineered controls than for procedural ones. For example, for a valve that automatically closes when pressure gets too high, we might say: it is protecting against overpressure in the vessel, its safety function is to shut off air supply and its performance standard is to be fully closed within 2 seconds of the pressure reaching a set level.
For procedural controls (and, in my view, it is perfectly fine to have procedures/activities as critical controls), you may have to get a bit more creative. For example, clipping on while working at height might have a safety function to ensure that harnesses are clipped to something adequately robust. Assurance checks would necessarily be visual confirmatory checks, but don’t forget they also have to extend to the availability of anchor points and confirmation that those are capable of taking the dynamic load of a falling person.

[Slight digression – this is a good example of making sure your critical control provides a full line of protection. It is no use providing engineering assessments of anchor points if nobody checks people use them (and vice versa) – your critical control should include ALL components to provide the safety function required. For our valve above, this would include detection as well as actuation.]
When this is clearly defined, assurance plans are easier to create, because we know what to test for. I usually advocate a two-tier regime. Tier 1 is a regular operational check and tier 2 more of a bigger picture, audit-type check. So, for our valve, the first check maybe a straightforward valve test – inject a signal and see if it closes quickly enough. The second check may be looking across all valves of that type and confirming that all tests were done, checking for any common themes in operational or test variability, and double-checking age profile of valves versus manufacturer’s recommended operating life.
Assurance plans may also include things like operator training required and, something that is often forgotten, what the substitution arrangements are if the control is unavailable and what action should be taken in the event of failure.
As for sharing with the Board, there are two components. Routine reporting should focus on effectiveness of critical controls and deep dives / awareness sessions should provide further insight into each of the critical risks and their controls. The Board need to be able to confidently answer the following four questions. If they can do this (positively), then both you and they have done your jobs successfully:
- Do I know what our critical risks are?
- Do I know how they are controlled?
- Are those controls effective?
- What am I basing that effectiveness decision on?
A quick summary template I use is as a starting point is shown below, but comprehensive good guidance for all of this is provided by the ICMM guide ICMM – Critical Control Management: Good Practice Guide
| Safety Critical Control Sheet | |
| Control description | What is the control in question? Include equipment number, procedure number etc if appropriate |
| Risk the control protects against | What risk(s) does this control protect against? |
| Safety function required | What is the exact function that the control needs to do to reduce the risk – eg a chlorine cylinder needs to maintain containment? |
| Performance standard | How well does it need to deliver that safety function – eg the chlorine cylinder needs to contain up to a certain pressure? |
| Proof test requirements | What test do we need to make to confirm the control will meet its performance standard and how often do we need to test? |
| Action to take in the event of failure | What if it fails – eg do we need to stop work immediately, just stop some activities or bring a standby on-line? |
| Action to take if the control is unavailable | If the control is unavailable what do we do – eg when we take a piece of equipment out for maintenance, can we swap it temporarily for something else, put a short-term administrative control in place, or simply not carry out the task? |
| Assurance check (administrative controls) | What do we need to do – audit/review/observation – to make sure the control is working? |
| Maintenance requirements (equipment) | What maintenance is required and how often? |
Next in this series on making health and safety governance practical, I will talk about influencing at a board and executive level. Make sure you don’t miss that, and future posts, and subscribe below (it’s free).
