So, you’ve decided to remove accident rates from your H&S reporting. Great idea – especially the mess that is lost time injury rates (see here). On making the proposal, you’re immediately met with the question, “So what will we replace them with?”
For some reason your preferred answer, “Nothing, they’re just not adding any value so let’s concentrate elsewhere,” doesn’t seem to cut it with CEOs and Boards.
There appear to be two reasons for this – firstly, they simply don’t like the idea of having less insight (this would be a valid concern if they were a source of genuine insight). Secondly, they are too conditioned by incident counting to move away from it (for which we, as their safety advisors, need to shoulder most of the blame). John Green once described this as being addicted to the stats. In which case, we need to wean them gradually off their addiction. Cold turkey is possible but is generally painful and often doesn’t work in the long run.
So, what to do?
I generally do one of two things (or both of them, or one after the other – all depends on your starting point).
- Take out the frequency rate graphs and replace them with a total number of incidents chart. Break these down by the potential consequence of the incident and then provide some narrative description of those that have the highest potential. This allows the focus to move away from the headline single metric onto good conversations about significant events, while still maintaining the feeling that incidents are being properly considered. You have to defend against the desire to trend by overall number and focus too strongly on that, but this is a helpful first step away from rate dependency.
- If you have a good handle on what your critical H&S risks are, you should be able to identify what the key controls are helping to manage those biggest risks. Then you can report on the effectiveness of those controls. This can take various forms – a simple count of how many of your controls are fully effective is easy but runs the risk of replacing one headline metric with another that is then blindly targeted. Although at least you know you’re only worrying about those risks that are genuinely of concern. Alternatively, report against progress on reviewing those key controls – are we keeping up progress on reviews and what are we learning each time. Reviews could be learning teams, audits, preventative maintenance checks or something else depending on the nature of the control in question. This, again, keeps the conversation focused on highest risk issues and begins the transition to thinking about how you demonstrate capacity to respond to unintended events by making sure your response is resilient – or even anti-fragile.
It can still be a long way to go to get them to really insightful discussions on complexity, variability, capacity and so on, but like all detox programmes – it’s one step at a time.
PS
This is particularly difficult if you are a contracting organisation and your clients demand these stats as part of considering your suitability for the job. In this case, you can continue to collect them quietly in the background and use them for this purpose only. At the same time, join up with some of your supply chain partners and start working with the clients to change the requirement. For some objective support – use this research paper The Statistical Invalidity of TRIR Updated 12.22.2020 (colorado.edu)