Posted By Frank Macleod
Colleagues
I am hoping you can give your opinions on the following two "schools of thought" regarding risk ratings when carrying out a risk assessment for any given task.

(1) the severity rating for a risk assessment never changes regardless of the control measures put in place although the likelihood will fall dramatically based on the same control measures. (e.g. if the severity is 5 and the likelihood is 5, the control measures are implemented and the likelihood drops to say 2. The severity stays the same as if the control measures fail the severity outcome would remain the same).

(2) once the control measures are put in place, both the severity and likelihood drop down ( i.e. the severity 5 and the likelihood 5 both drop as the control measures prevent the severity of injury damage etc as well as the likelihood).

Your thoughts will be appreciated.

Many thanks in advance.

Regards
Frank

Posted By holmezy

Frank,

continueing on from Chris,

severity can be reduced and should be a consideration. In a chemical environment using "chemical A" may be corrosive and therefore cause severe burns to the skin. However, if a control measure was to change to "chemical B" which may cause irritation, whilst still allowing the process to adequately function, then the severity has been decreased.

Think about hitting something with a small nail or pin with a 4lb lump hammer, yes you will drive the nail into the wood, but if you miss, you crush your fingers. If you substitute the lump hammer for a toffee hammer, you will still drive the nail, but if you miss, you recieve light bruising. Severity is decreased.

perhaps not the best examples, but think it shows what I'm getting at!!

holmezy

Posted By Tabs
You should remain flexible...

Falling off a four storey building would normally be fatal. Erecting a barrier will reduce the likelihood but not the consequence, so severity remains "fatal" or "high" if they fall over the barrier. BUT, if you choose a different control, such as big airbags below the fall, the consequence should change from "fatal" to "survivable", maybe minor injury. So severity could change to "Low".

Electrical shock can be fatal. Controls such as lock-out isolation, will reduce the likelihood, but if the worker makes contact with the supply, the consequence will remain "possibly fatal". However, work on live high voltage is possible if the right precautions are in place (no 'potential differences') so consequence goes from "possibly fatal" to "none" ... (this is a highly skilled and specialist task which must never be attempted by anyone who has not been thoroughly trained in the very few instances that it can be achieved - I am NOT advocating working on live equipment).

So... you have to look at each case and decide for yourself, "Has the consequential severity changed or not?"

Posted By Chris Packham
May I add one element to Merv's equation:

Severity x exposure x probability = risk

When doing a risk assessment we should always consider that situation which might arise should something go wrong. It is a bit late when that hose has ruptured and sprayed the worker with acid to start thinking about the risk. So with risk assessment we always need to be asking the "what if" question.

Chris

Posted By Andrew Meiklejohn
Good point Tabs

but i would always initially consider the hazards/risks without any controls as I'm sure we all know how easlly some control measures can be circumvented or not used.

The residual risk will then be affected by control measures which will include training supervision inspection maintenance etc.

Also, shouldn't the risk assessment include the risk of the airbags not being inflated properly?

Andrew

Posted By Andy Brazier
I don't think anyone has said this yet. A common mistake is to only consider one risk ranking, usually for the worst case consequence. The reality is that there are multiple scenarios and the worst risk may be for a lesser consequence that has a higher likelihood. This is especially the case when comparing before and after risk control measures because the impact of the controls may be different for different consequences.