Ian Long, Routledge, £29.99 paperback (ebook also available)

Simplicity in safety investigations


This is not a big book, but it packs a lot of ideas into 142 pages. The author, now a consultant but formerly in a senior OSH post at Australian miner and nickel refiner BHP Billiton, has a lot of experience to draw on but he is also clearly well read. One of the strengths of this book is how he harnesses theories from writers such as Todd Conklin and Daniel Kahneman to the service of accident analysis.




Long makes a virtue of this “recombinant innovation”, making new techniques by combining existing ones.

His strongest message, channelling Erik Hollnagel’s Safety II and Sidney Dekker’s safety differently approaches (see above), is that investigations do not have to be restricted to unpicking things that have gone wrong. Long’s recommended “outcome analysis” technique can be applied equally to a period with no recordable incidents as to a safety failure.

The basic investigation approach he advocates is a gap analysis between “work as done” – what was happening at the time of the accident, “work as normal” and “work as intended” – what the procedures or method statements prescribe. This can be applied to a small local investigation by the people involved in a task or a larger manager-led exercise after a serious accident or near-miss.

Data gathering should use a PEEPO structure, he argues, dividing information into the categories of people, environment, equipment, procedures/documentation and organisation.

He has sound advice about scene preservation, team formation and the attitude investigators should adopt: open-minded and curious, cultivating what he calls “generous listening” and using a coaching approach to draw information out of interviewees rather than closing down an inquiry with leading questions.

The jacket blurb suggests the book could be used by supervisors and managers as well as safety professionals. Long helpfully splits out the more detailed explanations of theoretical underpinnings such as “shared space” theory or the various heuristics that can bias investigators, into a section headed “The technical and scientific stuff”, which leaves a manageable 56 pages that could be passed on to a non-practitioner as a primer.

This is a well-written and well-edited book; for many readers used to a more functional approach it may not bring simplicity to their investigations, but it will surely add rigour.

Routledge (


Add new comment