Episode 4: ICAM and Investigating Beyond the Front Line

Theo Ashworth

Welcome to Getting to the Root. This is Episode 4, and today it's ICAM, the Incident Cause Analysis Method. I'll start with something we ran into back in Episode 1. With the 5 Whys, you can follow a chain of questions and still stop at the person who pressed the wrong button.

Nell Harding

Right, and that was the trap. The chain feels thorough, but if you stop at the operator, you've only explained the last thing that happened. ICAM is built to stop that from happening. It sets out to look past the front line and into the organisation, and it's used for accidents, incidents and near misses alike.

Theo Ashworth

Where does it come from?

Nell Harding

Australia, in the late 1990s. The sources we've got say it was developed by BHP, the mining company, around 1998. Professor James Reason helped, and so did the Australian Transport Safety Bureau, with input from safety representatives in various industries. People like Gerry Gibb and Dr Graham Edkins are named among the contributors. Edkins was Head of Human Factors at Qantas at the time, so aviation was in the room from the start.

Theo Ashworth

And Reason is the Swiss cheese man.

Nell Harding

He is. Picture a stack of cheese slices. Each slice is a defence: a guard, an alarm, a permit, a supervisor. In a perfect world every slice would be solid. In reality each one has holes, and the holes come from design, maintenance, training, supervision. Most days the holes don't line up, so nothing gets through. Occasionally they do, and there's a clear path straight through every defence.

Theo Ashworth

What I like is that the holes were there before anyone had a bad day. Reason's term is latent conditions. They sit dormant in the system, waiting. The person at the sharp end, the one who gets the blame, is just the one standing there when the holes line up.

Nell Harding

That's the shift. The older approach, the person model, asks what someone did or didn't do. One of the sources describes that as a superficial explanation. It describes the visible failures, the active ones, but it doesn't touch the context. So you fix the person, the same conditions are still there, and the next person walks into the same hole.

Theo Ashworth

I've been trying to think of an everyday version. Say a kitchen has a wet floor and someone slips with a hot pan. You could write it up as carelessness. But if the leak was reported three times, and the mats were never reordered because the budget was frozen, the cook is the last link in a long chain, not the cause of it.

Nell Harding

That's a hypothetical, but it's exactly the structure ICAM imposes. It's also why it caught on. It's been used in mining, oil and gas, heavy construction, rail, road, utilities and aviation. One of the training providers lists use in countries from Canada and Chile to Indonesia and South Africa, and says that suggests it travels well across cultures and industries. And Safety Wise says it's trained more than 70,000 people across 43 countries.

Theo Ashworth

Those are provider figures, so a bit of salesmanship is possible, but the spread across sectors is consistent across all three sources. So let's walk the layers.

Nell Harding

There are four, and we go from the sharp end to the blunt end. First: absent or failed defences. What should have stopped this and didn't? A guard that was missing or bypassed. An interlock that didn't work. An alarm that was silent. A permit, a competency check, supervision.

Theo Ashworth

Then the second layer, individual or team actions. This is what people did right before the event. The slips, lapses, mistakes, the deviations from procedure. And the key line is that ICAM treats these as symptoms to be explained, not causes to be punished.

Nell Harding

That's the line that changes the whole meeting. Instead of asking who did it, you ask what made that action make sense at the time. Which leads to the third layer: task and environmental conditions. Fatigue, time pressure, poor lighting, noise, weather, unclear procedures, awkward tools, confusing interfaces.

Theo Ashworth

And that's where I think most investigations could usefully stop and don't. You find the tired crew on a rushed job and you think, well, there it is. But ICAM says a tired, rushed crew didn't just happen. Somebody, somewhere, set that up. Which is the fourth layer: organisational factors.

Nell Harding

The upstream decisions. Contractor procurement policies, cuts to capital spending, production incentives that pull against safety, thin training budgets, how change is managed, the culture. Safety Wise describes sixteen organisational factor types in ICAM, so the analysis points at categories you can actually trend across incidents.

Theo Ashworth

Let me run a quick hypothetical, and I'm inventing this, it's not a real case. A maintenance worker bypasses an interlock on a machine and gets hurt. Layer one, the interlock was bypassed. Layer two, he did the bypass. Layer three, the job was behind schedule and the guard kept tripping on a poorly designed sensor. Layer four, maintenance spending was deferred and nobody owned the sensor problem.

Nell Harding

And notice the fix changes at each layer. Retrain the worker is a layer two answer. Replace the sensor and fund the maintenance backlog is a layer four answer, and that fixes it for every machine like it. The recommendations are meant to go that way, up the hierarchy of controls toward engineering and system fixes, not just retraining or disciplinary notices.

Theo Ashworth

How does a team actually collect all this?

Nell Harding

First secure the scene and gather evidence, often using PEEPO: people, environment, equipment, procedures, organisation. Then build a timeline, a sequence of events and conditions, not one moment. Then work each event back through the four layers. Some teams use 5 Whys inside that to test each causal link, though Safety Wise argues that's a BHP legacy habit and is redundant, because ICAM reaches the organisational factors anyway. They suggest 5 Whys, or a lighter ICAM, for low risk events, and full ICAM from moderate risk upward.

Theo Ashworth

So it's the next rung up from what we covered in Episode 1, not a rival.

Nell Harding

Right. And it doesn't need a disaster. It's equally applied to high potential near misses, where the last slice of cheese happened to hold. Those are free lessons. Same holes, nobody hurt.

Theo Ashworth

Which is the uncomfortable bit. If you only investigate harm, you wait for the holes to line up. The near miss is the system telling you for free.

Nell Harding

One caution, though. A report on a shelf changes nothing. The findings need owners, due dates and checks that the fix actually happened, and the risk behind the failed control should be reassessed.

Theo Ashworth

So next time something goes wrong, or nearly does, ask what the person did, then keep going. Ask what made it possible, and who decided that. That's the root. Thanks for listening to Getting to the Root.