A worker's hand is caught in a machine. The guard is refitted, the worker is warned, the file is closed. Eight months later, on a different shift, it happens again. This pattern is so common that many managers treat it as normal. It is not normal. It is a sign that the first investigation stopped too early.
The file closes at the first answer
Most internal investigations stop at the immediate cause: the missing guard, the wrong procedure, the careless moment. These are real, but they are symptoms. The question that was never asked is why the guard was off, why the procedure was not followed, why the moment was careless. Until that question is answered, the conditions that produced the accident are still in place.
The fix names a person
Retraining, warning letters and reminders are the most common corrective actions in the files we review, and the least effective. They change nothing about the production target, the tool, the staffing level or the design that made the shortcut necessary. The next person in the same job faces the same pressures.
Nobody checks whether the fix worked
Actions are marked closed when they are done, not when they have worked. A new procedure is issued; nobody checks whether it is followed. A barrier is installed; nobody checks whether it is bypassed. Without an effectiveness check, a closed action is just a closed file.
What stops the repeat
Trace every cause back to a condition the organisation controls: a target, a design, a staffing decision, a missing check. Write actions that change those conditions for everyone doing the job, not just the person involved. Then return after a few months and test whether the conditions have really changed. This is what we mean by fact-finding, not fault-finding.
KEY TAKEAWAYS
→Repeat accidents mean the first investigation stopped at a symptom.
→Actions that name a person rarely prevent recurrence.
→Test whether actions worked, not just whether they were done.
QUESTIONS READERS ASK
Why do accidents keep happening after investigation?
Because investigations often stop at the immediate cause and actions target individuals rather than the conditions that produced the accident.
How can we stop repeat accidents?
Trace causes to conditions the organisation controls, change those conditions, and check later that the changes worked.
