Now then fellow safety people

Wondering if anybody can assist me.

I am currently looking into setting targets for AFR(RIDDOR) rate and for LTIFR.

Business has approx 160 staff in a high risk enviroment and quite a lot of operational exposure

Cheers

Hi staggsy

What would you do with these targets?

Might sound like a stupid question but it isn't !!!

Notoriously difficult as what would you benchmark against?

Suppose your 160 people work in a factory making widgets. Some higher risk processes, some lower risk processes and you get X injuries of whatever severity in a typical year.

What do you read from getting 1.5X accidents the following year?

....or from 0.5X the following year?

Your numbers are probably too small to enable statistically significant change over a time period of less than SEVERAL years, perhaps longer.

There are SOOOO many variables that can change the numbers.

Examples - there is an economic downturn and 40 people get made redundant so all else being equal you would expect a 25% reduction in numbers of accidents, but it doesn't work like that as the remaining 120 work a little bit longer and those who are left might well be skewed towards the higher risk process.

.....or you get an order for fidgets - you bring in some antiquated machinery that you didn't need before but which is essential for the production of fidgets (but is useless for widgets) and suddenly your risk profile is perhaps skewed to higher risk activities and the number of accidents goes up - this does NOT necessarily tell you that performance is worse.

......or you outsource the higher risk processes and as long as you don't count supply chain accidents your numbers SHOULD go down. Are you performing better? Or are your sustainability credentials in tatters as it is now children working on unguarded machinery in some developing nation doing all the dangerous stuff?

Please don't go for ZERO!

Hi Steve

In most organisations I think setting ANY target for the reactive statistic of numbers of accidents or accident/incident rates is generally unhelpful and I certainly didn't offer the 1.6 you included in your posting.

History tells us that time after time when organisations set targets they result in presumably unintended consequences, such as underreporting and reclassification.

I used to work for Jacobs who lost 11 people at BP Texas City in March 2005, with nearly another 200 sustaining OSHA "recordables".

This happened shortly after BP received the Telos report. BP tried and failed to keep the report out of the public domain and it makes for difficult reading. 

The accident numbers were looking REALLY good. What was missing was any recognition of focus that needed to be given to the "low probability, high consequence" event that was waiting to happen. Telos predicted exactly the type of scenario that would kill 15 people.

.....as had HSE when publishing a report after three explosions at BP Grangemouth in 2000.

5 years after Texas City, OSHA was sufficiently fed up with the constant fiddling of the figures that they gave the onshore petrochem industry a sharp warning.

 “special attention to safety incentive and discipline programs that have been shown to discourage workers from reporting injuries and illnesses”

Now it is a standard principle of management that targets should be SMART and the A stands for Achievable and Zero isn't achievable and so it is not an appropriate target.

UK law generally requires what is reasonably practicable and we have almost entirely got rid of the "strict liability" duties in e.g. the Factories Act 1961 which the HSE didn't even try to enforce when that would have meant bringing UK plc to a standstill - HSE had literally 1000s of instructions to its Inspectors as to what to enforce INSTEAD and I very much doubt that a single Inspector had every one of these instructions in their memory.

Suppose you travel by train (may be you don't and use the much more dangerous option of driving everywhere), well that means that you have to stand on a station platform and on the balance of probabilities that means that there is a quantifiable risk of the statistical you falling from the edge of the platform and being injured to a lesser or greater (including fatal) severity when there isn't a train at the platform.

Most UK railway platforms do not have interlocked barriers to stop you falling from that open edge.

Now if you want ZERO then we need to retrofit lots of interlocked barriers AND we would need to make sure that all the doors on the carriages align with the interlocked access gates AND we would need to make sure that trains all stop within a very small tolerance so that the train doors align with the access gates.

But would that be a sensible way forward or should we accept some level of risk?

....and if you don't use the train or drive, surely you walk on and cross the road?

In the big bad world outside work or in the workplace setting, people can and will slip and trip on the level however assiduous management and workforce (or others) are at keeping footways in good condition, free of debris and spillages. There is a small risk and that means a small risk of accidents and injuries AND however good the traffic management there is a small risk of vehicle hitting pedestrian. So going for ZERO is not attainable.

Then there is also the elephant in the room.

All these organisations spending hours on producing and reviewing accident statistics and pretty graphs purporting to show declining trends in numbers or rates.

But, the same organisations rarely predicting how many people (or the environment) will suffer perhaps fatally as a result of exposure to risks that may not have immediate impacts.

...and if the ILO estimate that 2.6 million people die prematurely each year around the World as a result of work-related causes and that over 2m of those die from occupational disease, (+ similar proportions for less severe outcomes) with such numbers being likely to translate broadly at the micro level of the individual organisation, why is any organisation more interested in this year's accidents that the occupational ill health coming 5, 10 or 20 years time?

Can I ask SteveL a Question: are you saying that your H&S management system is so good that you can guarantee that no accidents will happen or do concede that (reportable?) accidents might and you are to a certain extent lucky that no such accidents have taken place?

And of course what would happen if an accident WAS to occur? What would the consequences be?

Achieved absolute zero or merely zero reported?

Not one paper cut, stubbed toe, stapled finger, hammered thumb, banged head, door slam, trapped finger.

If an incident happens and no one reports did it occur?

Originally Posted by: A Kurdziel

Can I ask SteveL a Question: are you saying that your H&S management system is so good that you can guarantee that no accidents will happen or do concede that (reportable?) accidents might and you are to a certain extent lucky that no such accidents have taken place?

And of course what would happen if an accident WAS to occur? What would the consequences be?

No I am not saying it is so good that I can guarantee that an accident will not happen, and yes we might be lucky, but trying to make it work and getting the support from bottom up helps. 

And is an accident does occur, try an  dfind out how to prevent happening agian and provide all the information to all the workforce & management. 

As to the consequences, with regards to what ? the severity to the injured person, or the fact that we have to say we had somebody who was hurt and we have to try and prevent it, why because we care. 

Well the answer to that ethical difficulty is instead of saying your target =2, saying your target is <3.

The cynic in me has come to the fore and so I'll share my thoughts for what they are worth.

If I was able to predict what accidents would happen, knew without doubt the root causes, had the finances and other resources to tackle these root causes ... then I might be able to stomach 'Zero' as a target. Since random chance, dumb luck or fate plays a much bigger role than we all admit, I for one am certain that whatever I do, however hard I work, no matter how much buy-in I get from all and sundry that I can't prevent all accidents from happening .. so backward-looking accident statistics are pretty useless. Also as others have said measuring the differences between small numbers and asigning any significance to them is a mug's game