An organisation learns about its own risks in two very different ways, and most organisations are quietly excellent at one of them and have barely tried the other.
The first way begins after something goes wrong. An event occurs, someone is harmed or nearly so, and the organisation investigates: what happened, why, and how do we stop it recurring. This is retrospective learning, and its classic instrument is Root Cause Analysis. It is indispensable. A hospital that does not learn from its incidents is condemned to repeat them, and doing this work well is a mark of a serious institution. But it has three limits built into its nature. It needs a casualty before it produces a lesson. It examines the single pathway that actually failed, while the process may hold a dozen others that did not fail this time. And it works under the long shadow of hindsight, where the chain of events always looks more obvious after the fact than it ever did in the moment.
The second way begins before anything goes wrong at all. A team examines a process while it is still working, or appears to be, and asks where it could fail, how badly, how often, and whether the failure would be caught in time. This is proactive learning, and Failure Mode and Effects Analysis is its most developed form.
The comparison that sticks with most teams is driving at night. Root Cause Analysis is the rear-view mirror. You need it, you should check it often, and it tells you with great clarity what you have already hit. Proactive analysis is the headlights. It is the only equipment on the vehicle that shows you what is coming while you still have room to steer. A driver who navigated solely by the mirror would be considered reckless, and yet a great many safety programmes are built almost entirely on it, filing their corrective actions diligently and meeting the next foreseeable failure with the same surprise as the last.
I am not arguing that the mirror should be cracked or ignored. Retrospective and proactive work reinforce each other, and the best moment to run a proactive analysis is often on the redesigned process that an investigation has just produced. The argument is about balance. If everything your organisation does about safety is triggered by harm that has already happened, you have decided, in effect, that patients will supply your research questions, one incident at a time. There is a cheaper and kinder way to find the questions.
Turning the headlights on does not mean turning the mirror off. It means admitting that a device pointed permanently at the past is a strange thing to steer by, and that the road ahead has been dark for no better reason than that nobody reached for the switch.
Before Harm Happens is, in the end, an extended argument for reaching for the switch, and a practical guide to what the headlights show once you do.
This article accompanies the book Before Harm Happens.