Hi,

I am looking to introduce the Bradley Model to allow benchmarking our safety culture within the business. Having spent time looking for a basic questionniare to roll out to staff, I seemed to have fallen short.

I am just wondering if anyone had something they were willing to share or point me in the right direction.

Thank you 

Edited by user 06 February 2024 10:14:57(UTC)  | Reason: Spelling

Hi Les

Following on what AK has written, how would you intend to "benchmark"?

What would you use as your criteria for benchmarking and who would you choose to benchmark against?

What is usually done is to look at the Lost Time Injury Frequency Rate (LTIFR) or some variant such as rates of SIFs (Serious Injuries and Fatalities) or SIIFs (Serious Injuries, Illnesses and Fatalities) or more recently PSIFs (Potential Serious Injuries and Fatalities) but all are likely to lead to under or even overreporting.

PLUS it is very difficult to detect changes to your numbers or rates that are statistically significant unless you are either a mega organisation or look at a period of usually several years or more.

PLUS it is very difficult to find organisations that to benchmark with that are sufficient similar to yours. As example, if you compare two large Conractors one could have a policy of outsourcing most of the higher risk activities and then not counting the statistics for the supply chain. The other might wish to keep as much as possible in house and thence SHOULD have higher numbers or rates all else being equal.

Then as soon as you set targets you face problems.

Let's say you take PSIFs. If you set a low target, then people will be influenced into NOT reporting the near misses (that you could learn so much from).

Conversely if you set a high target, you are encouraging the reporting of the minor issues that previously the supervisor or worker would just sort out there and then and not be bothered with bureaucratic reporting. The spill can and will happen. Do you learn that much from a report saying that the spill has been cleaned up UNLESS you have enough reports to detect some underlying reason for SOOOO MANY spills?

.....and if you use any of the usual models you are probably not going to measure the harm that is coming in the future whether that be chronic ill health or impacts on the environment OR the "low probability, high consequence" scenario.

DuPont used its Bradley Curve as a major selling point for its H&S consultancy - but just as others have celebrated what looked like very good LTIFRs, the disaster followed - a quadruple fatality.

The "others" include BP - Grangemouth 2000 (three close calls), Texas City 2010 (15 dead and nearly 200 injured) and Deepwater Horizon and Vale at Fundao 2015 followed by Brumadinho.