Stop Investigating Everything: How Smarter Event Classification Reduces Risk

A worker trips over their own feet on a flat walkway at 10am. Boots fine, not fatigued, not rushing. They fracture an arm. Under a traditional classification system, that event triggers a month-long investigation,  which concludes with a toolbox talk telling people to watch where they're walking. Sound familiar?

In the opening session of the Investigation Series, Mark Alston of Investigations Differently made a deceptively simple argument: we investigate far too much, far too uniformly, and we learn far too little for the effort. This article distils his case for classifying events smarter, so your team's energy lands where it actually reduces risk.

We investigate everything the same way, and it's not working

Australia loses around 200 workers a year, a figure that hasn't shifted in two decades, even as we file more reports, run more investigations and close out more actions than ever. Alston's explanation is uncomfortable: when a paper cut and a near-fatality get the same process, we overload our teams and lose sight of the events that carry serious risk.

He points to real examples, an auto dump truck that drove over its operator's own parked ute (no credible potential for a fatality, yet flagged as high-potential), and a single mine that ran 245 ICAM investigations in one year. Effort, but not learning.

“Toolbox talks have never fixed a thing in their life. This has got to stop.”

Our injury classifications are largely made up

Alston's more provocative point is that the definitions underpinning most classification, first aid, medically treated, restricted, have no solid basis. The standard many were built on stopped being published years ago, and neither the model WHS laws nor ISO 45004 provide injury definitions. In practice, organisations invented their own, sometimes tuned to minimise recordable injury rates.

The consequence is statistical manipulation and under-reporting, a focus on high-frequency, low-consequence events, and investigations driven by a risk matrix rather than by any genuine opportunity to learn.

A better question: what will we actually learn?

The fix isn't more effort, it's a clearer objective. Instead of the unachievable goal of ‘preventing recurrence’, Alston argues investigations should exist to learn about and reduce risk. That reframes everything, because risk is the effect of uncertainty, the things we don't yet know or understand.

“What have we learned that we did not know before the investigation? If the answer is nothing, why did we bother?”

His Event Learning Assessment White Paper (co-authored with Jop Pavlinga at Griffith University's Safety Innovation Lab) turns this into a practical decision framework. Was there credible potential for a critical event? Is it notifiable and is there an opportunity to learn? Was critical control absent or failed? Depending on the answers, you land on a full investigation, a learning team, a lighter ‘event insight’, or simply recording it and moving on.

What this means in practice

Two ideas make the framework work day to day. First, ‘credible potential’ means reasonable and plausible, not every conceivable what-if. Tripping in a car park could, in theory, kill you; it isn't credible that it will. Second, and hardest culturally, you must give your team ‘permission not to investigate’ what won’t teach you anything. Not every event carries a return on the investment of a full investigation.

For organisations using myosh, this is exactly the shift the platform is built to support, capturing richer, narrative incident data and helping teams focus effort on the events that genuinely reduce risk, rather than drowning in uniform paperwork.

Key takeaways

• Investigating every incident the same way overloads teams and hides serious risk.

• Most injury classifications lack a solid evidence base , don't let the risk matrix drive the investigation.

• Set a clear objective: investigations exist to learn about and reduce risk.

• Use ‘credible potential’, reasonable and plausible, not worst-case what-ifs.

• Give your team permission not to investigate what won't teach them anything.

• Every investigation should deliver a return: new understanding you didn't have before.

‍

Watch the full session

This article only scratches the surface of a rich, example-packed hour. Watch Mark Alston's full session on demand to see the Event Learning Assessment in action.  

This session is part of the myosh Academy free webinar series. Explore the full library and register for upcoming sessions in the Investigation Series.

Presenter: Mark Alston, Founder, Investigations Differently.

Presenter: Mark Alston, Founder, Investigations Differently.

Tags/Categories: Workplace Safety, HSEQ, WHS, Incident Investigation, Risk Management.

‍

Internal links: ‘myosh Academy webinars’ → https://www.myosh.com/academy/webinars ; related — the Critical Control Management module. Suggested image alt text: ‘Safety professional reviewing an incident classification decision framework’.

‍