Hi,

Within the term residual risks with risk assessment i've recently had a query concerning the severity rating so though I'd gauge opinion, If you put controls in place does the severity stay the same as before the controls or lower? I.E. if working on a roof the severity of falling is death. If you use mansafe system and harness is the severity still the same? The severity could be deemed as lower as the controls if followed lower it and the likelihood is lower or is the severity the same but the likelihood lower?

But then under CDM if a i have roof and no edge protection there is a residual risk left but if I  put a parapet round it would i put in my residual risk regitser roof- potential fall- or has the risk been designed out?

Thanks, 

Thanks Chris

My kind of issue is that if the controls are followed how can the severity be the same? If the the controls, or one of the controls are taken out, then the severity raises as well as the likelihood.

If you had a roof with the edge protection would the initial assessment say severity of high for falling? or would you put it as low as you couldn't fall?

Thanks

Peter,

I am not sure I agree with your example of asbestos. Evidence shows that one exposure can be sufficient for a fatal outcome, but that the effect can be chronic, i.e. not become apparent for many years. More frequent or concentrated exposure may - but not necessarily will - increase the probability (i.e. the risk), but the outcome will be no different. So for me the severity remains the same, only the timing is different. 

A similar situation can exist with the possibilty of developing an allergic contact dermatitis. Sensitisation can occur on first exposure, or it may be only after repeated exposures, possibly at long intervals and over a long period. Sensitisation is the precondition for an allergic reaction that may occur immediately after the sensitisation (which itself is asymptomatic and generally permanent) or only much later (if ever) even though there has been repeated exposures in the interim. As the severity of the reaction will be similar in either situation I will argue that it does not change, only the potential for it to happen.

Methinks you are confusing terms and ideas so possibly also yourself.

Working on an unguarded roof presents a hazard.

The outcome of conducting a task in this location could be nothing at all or death through the sudden stop at the end of a major fall.

We can put in a control of no roof working - the outcome severity remains the same (death) but our residual risk is lower as the likelhood of exposure is eliminated.

We can put in a control of fall bags - the outcome severity is reduced from death but our residual risk may be the same as above as our exposure has now increased.

We can put in a control of harness / mansafe - the outcome severity may be the same as with no control measure as in this scenario we have switched from the sudden stop to suspension trauma so the residual risk may remain the same as with no control measure.

In this latter scenario we now identify a new hazard arising from our intervention - if we put in place measures to reduce suspension e.g. descent devices or a system of urgent rescue then the severity can be reduced and thereby the residual risk.

Thanks All

I have had a client saying when using a fall restraint harness and mansafe system the severity is the same as not using the controls.

if one of the controls was taken away i'd say the severity goes back up but if all controls employed it has to come down. He thinks not

Thanks

Slightly tangent. But you can reduce severity using controls. A good example of this relates to window cleaners. It used to be one of the most dangerous jobs. Now they do all their cleaning from ground level. The severity has been reduced from death to something lower.

As for the asbestos issue. I thought we had all moved away from the "1 fibre can kill you" days.

http://www.acsrisk.com/2017/12/roger-s-thoughts-breathe-a-single-asbestos-fibre-won-t-kill-you/

"Some years ago, Professor Seaton’s early post-mortem studies at the University of Aberdeen indicated that more than 60% of people in the UK had asbestos fibres in their lungs at the point of death and it was nothing to do with the cause of death. There are over some 60 million people in the UK. If 60% of them have asbestos fibres in their lungs then some 36 million people in the UK have asbestos fibres in their lungs. If one fibre kills, then all of these people would die of the asbestos-related disease. On average it takes about 30 years from overexposure to death which means we would expect some 1 200 000 asbestos-related deaths per annum. As the real figure of asbestos-related deaths is some 4,000 pa, it seems very reasonable to say that no, one single fibre of asbestos does not kill."

Quote

But the residual risks are not what are left after the controls are in place? If I have a floor plate on a building with scaffold around is the likelihood falling from the floor? If i then install glazing full height is the severity still falling?

Once the controls are in place and followed I would have thought, maybe wringly the severity comes down.

End Quote

The severity is for example how badly someone can be hurt.  Regarless of control measures if you fall from a given height the potential injury would be the same.

The residual risk is the adjustment given the new likelihood.  The only way you can reduce the severity is if you do something to reduce it, in the now less likely event a person falls you have a control measure that lessens the impact when they hit the ground.  However, if you have a control that eliminates the fall, you wouldnt need a control to reduce the severity (there would be no point because no one can fall).

Think of it the other way…..The fall protection reduces the likelihood of someone falling and hitting the ground (assuming they use that control measure).  Airbags (for example) might reduce the severity if there were no other controls in place.   In theory this would reduce the overall risk level the if airbags were the only control measure the likelihood would still be the same but the severity lower.

What I am saying is that you would need to have controls in place to affect both the likelihood and the severity if both were to reduce…..but it's not alway nescessary

Edited by user 16 March 2020 15:22:50(UTC)  | Reason: Not specified

Is there some cross-purposes going on here?

The OP referred to residual risk and CDM.  The example given relates to design risk management.  Most of the responses have been about RAMS-type risk assessment (i.e. assessing a work operation).  Residual risk in CDM designer risk management is rather a different thing.

DRM under CDM is even less suited to the 5x5 matrix than are RAMS RAs.

In the CDM context, residual risk is just the risks that  you haven't managed to design out, so need to communicate to the people executing the works.  You don't rank or rate or score these, you point out they  exist, and preferably identify how you anticipate they might be managed.  They might be residual risks because you managed to reduce something to a lesser risk, but it's still significant, or it might be something that you just can't see a way to eliminate.

The CDM comment appears secondary as it discusses a parapet without clarification of the height of such installation i.e. how much protection would be afforded.

Thanks

I put the CDM comment in to show a parralel and interpretation of residual risk.

In terms of Risk assessment i agree with a lot of the comments that if the controls are in place the severity is lower. I agree if they fail the severity rises but if they can fail then the risk assessment probably isn't suitable and sufficient in the first place. Yes human eror can take place but that doesn't form part of all the controls in the first place.

Going back to the flat roof. If a mansafe- tested and in date- and harness and fall restraint lanyard- and fully trained operatives the seveirty in my view must be lower as its the risk in place with the controls attached to it. You are not starting back at point A of working on a flat roof with no protection.

Hi Joan

With respect, I think you are still confused with this concept - the SEVERITY does not change. It is the RISK that changes when controls are added. For example, falling from a roof with or without controls the SEVERITY stays the same. 

The vagaries with the English language means that words can be used intechangibly to mean much the same thing which can cause confusion. So a hazard could also be called a risk, and so on.

Hazard = tiger

Risk =it will bite you

Likelihood = it will bite you is high

Severity = high

Control option 1 =replace with less dangerous animal = Kitten

Remaining Likelihood = still high

Remaining severity = Low

Control option 2 = put tiger in cage

Remaining Likelihood = medium as if you put arm in cage it will still bite you

Remaining severity = medium as it is less likely it will be fatal

Control option 2a = as well as cage, put a barrier well over arm’s length away from cage all around.

Remaining Likelihood = Low

Remaining severity = Low

Control option 3 = find Vegan tiger

Remaining Likelihood = high - no such thing

Remaining severity = high - no such thing so not practical control

When doing the risk assessment, you are saying what the controls will do for you when used. It will be up to others to enforce. If controls not put in place or inadequate or not practical then original risk remains.

Chris

I agree with Chris42's post. Looking at this thread has me questioning what I have been taught regarding "severity"

Surely severity is the result of the "accident". For example death is very severe, a scratch is not. If you introduce a control that stops you being able to die, but still leaves the residual risk of getting a scratch, have you not reduced the severity of the risk?

So going back to working at height, if you can no longer fall from height, when working from the ground, for example, then haven't you reduced the severity from death to a lower state?

Chris

Maybe I am missing something. In your first post you say the "remaining severity" is reduced to medium or low with engineered controls. In your second post you say the engineered controls reduce the likelihood but not the severity.

I guess it is you use of "remaining" that is not clear.

Regarding substitution. If the hazard is "tiger" I really don't think you can claim replacing with a kitten to be the same scenario. I did cover this in my first post that changing the hazard actually changes the activity. Great risk management but not a good illustration of risk assessment.

Unfortunately we continually get tripped up by use of terminology. If we apply it correctly and consistently we can have a good debate on actual risk. Because we tend to mix it up everyone gets confused.

Originally Posted by: andybz

Chris

In your second post you say the engineered controls reduce the likelihood, but not the severity.

Fair point, but your problem was with my first post. In my first post As I noted there was still a potential severity as an arm could be put in the cage. In the option 2a I considered the person outside of the 95 percentile norm, with extra-long arms.

In my second post I generalised that engineering controls often only alter the likelihood, ie machine guarding. However not that hard and fast, and still feel mu submission shows how both Likelihood and severity can change, not just likelihood.

I stick with substitution being a valid control on the assessment being done. Perhaps If I gave the assessment a title it would help. I will go with "Owning an animal”

Chris

I think the problem with the way some or many risk assessments are done is the control measures typically only try to reduce the likelihood of something happening or only try to reduce the severity (PPE type controls) where they really should try to do both.  Only in an instance where you 100% eliminate the hazard should not having a control in place to reduce the severity be acceptable.  Often a reduction in the liklihood of the event occuring is relied upon as effective where the severity would remain the same if that reduction in liklihood control failed or was ineffective.

A tiger attacking someone who puts their arm in the cage could still very much kill that person as could the bite from a house cat and in any case lets face it most of the time we need the tiger for the job, the kitten just won’t cut it so you still need controls to reduce the severity of the bite.

Falling from a roof with a harness on can still kill you if you are left there for long enough therefore you have a control in place to reduce that severity hence the requirement for a rescue plan etc.  Designing the hazard out by installing a parapet eliminates the likelihood of the event occurring therefore addressing the severity is pointless.

Controls that reduce the likelihood (not eliminate) should have further controls that reduce the severity unless those reduction in liklihood controls can demonstrably do both, which is not always the case.  And unless you can render a person completely unharmed by reducing the severity to a negligable level when something bad happens there should be some control in place to reduce the liklihood of the event happening.

Edited by user 18 March 2020 10:03:40(UTC)  | Reason: Not specified