Rank: Forum user
|
Been asked to look into how to calculate Accident Frequency / rates. Company I work at, employ 95 people on one site, and 30 on another. There is a mix of Office and Warehouse Staff. Getting confused
AFR = No of lost time accidents x 100,000/No of man hours worked
AIR = No of work related injuries x 1000/Average No of persons employed
Are these formulaes correct....I've read others that have AFR calculating RIDDOR's and not lost time (by lost time, is it meant days or hours
|
|
|
|
Rank: Forum user
|
example ive worked out for AFI
We had 25 reportable injuries in 2014. 36,000 Hours were worked roughly
25 x 100,000 divide by 36000 = 69
So if right...Over course of year, we would have a rate of 69..that seems very high
|
|
|
|
Rank: Super forum user
|
Jon Joe - when you refer to "reportable injuries" do you mean RIDDOR reportable or are you meaning lost time injuries (how ever you have counted those)?
|
|
|
|
Rank: Super forum user
|
Have you got your hours right ? Sounds like you have about 17 employees ?
Chris
|
|
|
|
Rank: Forum user
|
Chris42 wrote:Have you got your hours right ? Sounds like you have about 17 employees ?
Chris it was just an example, to show the calculations are right....hours worked was just a random number
|
|
|
|
Rank: Forum user
|
peter gotch wrote:Jon Joe - when you refer to "reportable injuries" do you mean RIDDOR reportable or are you meaning lost time injuries (how ever you have counted those)? see, this is where im getting confused...correct me if im wrong.....If I want to work out AFI, I am basically working out RIDDOR's??? If I want to work out all Injuries in workplace, im working out AFI???
|
|
|
|
Rank: Super forum user
|
You can work out both with the same formula ie
RIDDOR AFR you just use the RIDDOR accidents or normal AFR you use all lost time accidents
so 25 lost time accidents of which 4 are RIDDOR reportable :-
RIDDOR AFR is 4 x 100,000 / 36000
or
Lost time AFR is 25 x 100,000 /36000
You just need to qualify what information you are presenting.
|
|
|
|
Rank: Forum user
|
last question on the subject.....if im doing a 12 month rolling rate for accidents per person...if the employee numbers are different by month, do I include an average number of staff or the number currently?
|
|
|
|
Rank: Super forum user
|
When I was doing accident rates I did it by the number of people working per month and the actual hours worked. It did make it more time consuming but also did make it more accurate. One of my predecessors used an excel formula which automatically assumed everyone worked 37.5 hours a week, which almost 75% of the workforce didn't, so all of a sudden there was panic because the accident stats went up- there was nothing wrong, they were just a bit more specific.
|
|
|
|
Rank: Super forum user
|
A quick comment for organisations that don't actually measure hours worked - there used to be lots with 'time clocks', but that's pretty rare today!
In round numbers, each person works ~2000 hours a year. So 100 people (a group size most people can envisage) = 200,000 hours/y roughly. That's the same number as US-based organisations use to calculate injury rates, whereas the UK tradition is per 100,000 hours. So when comparing rates, you have to be careful about which calculation method is used.
However, HSE gave up quoting injury rates per hours worked years ago - now they use per 100,000 full time employees. Such numbers may be OK for comparing sectors, but they don't make much sense for a typical organisation, or part of it. So IOSH recommend (in our guidance on reporting) using per 100 employees, as the numbers need to be 'manager friendly' to be of much use within the organisation and, as noted above, most managers and others can visualise a group of 100 people, but havn't a clue what 100,000 or 200,000 hours looks like! If you then want to compare your reportable injury rate with HSE data, just divide their figures by 1000 - most of them are of that order anyway.
Final point: there will always be some random variations, so it's of little meaning to quote injury rates to several places of decimals, which is also why there's not much value in using HR resources or others to count worked hours if the data is used only to calculate injury rates and for no other business purpose. If the counting is done for other business reasons, then of course use the most accurate data - but does anyone at the top actually care whether we quote rates per hour or per person, as long as we do it consistently?
|
|
|
|
Rank: New forum user
|
A quick question - is it possible to compare the accident rates of organisations which use different normalisation's; I.E US based firms which use 200,000 & UK firms which mostly use the 100,000 hour normalisation?
|
|
|
|
Rank: Super forum user
|
labrat23 wrote:A quick question - is it possible to compare the accident rates of organisations which use different normalisation's; I.E US based firms which use 200,000 & UK firms which mostly use the 100,000 hour normalisation? It's just a rate. It doesn't matter what units you use. 50 miles per hour is the same thing as 80.5 kilometres per hour is the same thing as 134,400 furlongs per fortnight. The unit doesn't change the speed. 10 accidents per 100,000 working hours is the same thing as 20 per 200,000 working hours or 1 per 10,000 hours. What is significant, however, is the definition of accident. That's what makes comparing rates between different jurisdictions difficult. Fatals is OK - most places use a similar definition of 'dead', though even then there are differing definitions of what constitutes at work. Injuries is much more difficult - is it three day, five day, seven day, any lost time, professional medical treatment required, any treatment required? Are you comparing anything from paper cuts up in one jurisdiction against 7-day lost time in another?
|
|
|
|
Rank: Super forum user
|
quote=labrat23]A quick question - is it possible to compare the accident rates of organisations which use different normalisation's; I.E US based firms which use 200,000 & UK firms which mostly use the 100,000 hour normalisation? Yes if you first either multiply the UK figure by 2 and you can compare direct to US number or Divide the US number by 2 and you can compare to UK number directly. Chris
|
|
|
|
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.