Executives, regulators and crews read investigation reports without having seen the site. What to define, what to show, and why a definitions table earns its place on the page.
The people who most need an investigation report are the people least equipped to read it. An executive deciding whether to fund a fleet-wide change, a regulator forming a view of your systems, a supervisor on another site wondering whether this could happen to their crew: none of them were there, none of them sat through the interviews, and most of them do not work in safety.
Investigators write for each other by default. It is a reasonable instinct, because the immediate audience is the review panel. It is also how a good investigation ends up with no effect outside the team that ran it.
ICAM has a working vocabulary, and it is precise for a reason. DF for absent or failed defences, IT for individual and team actions, TE for task and environmental conditions with the human performance factors inside it, OFT for the organisational factor types. Work as imagined, work as normal, work as done. A reader who has done the training moves through that comfortably. A chief financial officer does not, and will either skim it or quietly decide the report is not for them.
The answer is not to strip the vocabulary out. Removing it flattens the analysis and makes the report harder to defend later. The answer is to define the terms in the report itself, once, in plain language, near the front. A short definitions table costs half a page, and it is the difference between a report that circulates and a report that gets filed.
Write those definitions for somebody outside the safety function. "Controls that were missing or did not work on the day" does more work than a textbook definition of a defence.
Terms of reference are usually treated as a formality, and they are the most reread section in any contested report. State the event plainly, state the period examined, and state what the investigation deliberately did not cover. A boundary that is written down is a boundary you can stand behind. A boundary that is only implied reads, months later, as something you missed.
A report that records only the actual outcome invites the reader to size the response to the injury. An operator who fell and was bruised generates a bruise-sized response, even when the same fall a metre to the left is a fatality.
Carry both: the actual outcome, and the credible worst outcome with the reasoning for it stated. That pairing is what lets a reader who was not there understand why a comparatively minor event is being treated seriously, and it is the section that most often changes what gets funded.
A recommendation on its own asks to be taken on trust. A recommendation attached to the factors it answers, and those factors attached to the evidence behind them, can be checked by somebody who was not in the room. That traceability is what makes a report defensible under scrutiny, and it is also what stops a good recommendation being quietly dropped in budget season, because the cost of dropping it is visible.
SafetyPulse assembles the report from what the investigation already holds, including the definitions, the terms of reference and the severity classification, so the structure is there before anybody starts writing. The investigator still edits and approves every line of it. The point is that the reader who was not there gets a document built for them, rather than a document built for the panel and forwarded on.
SafetyPulse by ICAM Australia Pty Ltd. ICAM training and methodology.