If an investigation report lands on my desk with “human error” listed as the root cause and “retrain the employee” listed as the corrective action, I am sending it back.
Not because the employee’s actions do not matter. They do. If someone disconnected a lanyard, removed a guard, or ignored a spotter, the report should say so plainly.
But that only tells me the last thing a person did before the incident. It does not tell me why the incident happened.
After more than 36 years in construction, I have seen the same routine play out too many times. An employee gets hurt. Everyone meets in the job trailer. Before the interviews are finished or the equipment has been examined, someone says, “He knew better.”
That may be true. It still is not a root cause.
A few days later, the employee sits through another class and signs another training sheet. The company can show that something was done, but the jobsite, equipment, supervision, and work process remain exactly as they were before the injury.
That is not prevention. It is paperwork.
Why Companies Stop at Human Error
I understand why companies use human error as the answer. It is simple. It puts the problem on one person, gives management an easy corrective action, and avoids uncomfortable questions about planning, equipment, production expectations, or supervision.
The trouble is that “human error” tells us almost nothing.
Did the employee forget a step? Did he misunderstand the requirement? Was the correct equipment unavailable? Was the written procedure impossible to follow under the actual conditions? Had the crew performed the work the same way for months? Did the foreman know about it? Was the employee trying to meet a schedule using the only method that appeared available?
Those are very different situations. Calling all of them human error hides the differences instead of investigating them.
OSHA’s own investigation guidance warns employers not to stop at carelessness or failure to follow a procedure. It tells investigators to look at the management, design, planning, organizational, and operational conditions behind the event. OSHA and EPA also recognize that incidents often have more than one root cause. OSHA Incident Investigation Guide
Documenting the employee’s action is necessary. Treating it as the final conclusion is not. What the investigator does next separates an investigation from a blame exercise.
Look at the Work, Not Just the Last Action
Take a worker transitioning from an extension ladder onto a roof. He disconnects his lanyard during the transition and falls.
It would be easy to write, “Employee failed to maintain 100 percent tie-off.” That statement may be factually correct. It still does not explain the incident.
Could the employee reach the roof anchor while standing on the ladder? Was another anchorage method available? Did the pretask plan explain how the transition would be made? Had the foreman watched other employees disconnect at the same point? Was the crew told to begin roofing before anyone verified the access method?
Suppose the investigation finds that the anchor was too far from the ladder, the crew did not have the connector needed for the transition, and the pretask plan simply said “100 percent tie-off” without explaining how to achieve it.
At that point, telling the employee to remain tied off adds nothing. He already knew the rule. What he did not have was a workable way to follow it.
That does not automatically excuse his decision. He could have stopped and asked for another method. His foreman could have stopped the work. The superintendent could have verified access before the crew was released. There may be responsibility at several levels.
That is the point. The injury did not come from one decision made in isolation.
I use a simple test when reviewing these reports: If another qualified employee steps into the same conditions tomorrow, have we changed anything that would keep the incident from happening again?
If the answer is no, we have not reached the root causes.
Retraining Must Correct a Training Problem
MSC Safety Solutions is a training company, so I am certainly not arguing that training is unimportant. Good training is essential. Retraining can also be required and entirely appropriate.
But training has a specific job. It develops knowledge, understanding, and skill. It does not repair every failure in an organization.
Before writing “retrain the employee” on an investigation report, finish this sentence:
We are retraining this employee because the investigation proved that he did not know or could not demonstrate ____________________.
If the blank cannot be filled with a specific knowledge or skill deficiency, retraining may not address the cause.
An employee who was never trained on a task needs training. An employee who cannot demonstrate how to inspect or use the equipment may need retraining. If the equipment, procedure, or workplace changed and the old training no longer applies, retraining may be necessary.
OSHA uses that same basic logic in specific construction standards. Under the fall protection training requirements, retraining is required when changed conditions make earlier training obsolete or when an employee’s knowledge or use shows that the required understanding or skill has not been retained. OSHA’s scaffold standard contains similar language. 29 CFR 1926.503(c) and 29 CFR 1926.454(c)
Notice what those requirements focus on: understanding and skill. They do not say every injured employee must automatically be sent back through training.
Sometimes retraining is little more than discipline with a sign-in sheet. The company wants a documented response, so the employee is told to repeat a class whether or not the investigation found a training problem.
That may close the action item. It does not necessarily reduce the hazard.
A Class Cannot Fix the Wrong Problem
A toolbox talk cannot install a missing guardrail. A fall protection class cannot move an anchor within reach. Another grinder lesson cannot make the wrong tool suitable for a restricted space. A reminder about situational awareness cannot separate pedestrians from backing equipment.
If the procedure cannot be followed in the field, rewrite the procedure or change the work. If the correct equipment was not available, fix procurement and mobilization. If supervisors have tolerated the same shortcut for six months, address supervision and the expectations being communicated to the crew.
Training may be part of those corrections, but it should not be used to avoid them.
The NIOSH Hierarchy of Controls places training among administrative controls. Administrative controls depend on people consistently remembering and applying them. Whenever practical, stronger controls remove the hazard, separate the worker from it, or make the wrong action harder to perform.
This connects directly to a larger problem in our industry. We often confuse training with workforce development. Training gives people information and skills. Workforce development also gives them competent supervision, the right tools, realistic planning, clear authority, and the judgment to stop when the plan no longer fits the work. I addressed that difference in Why OSHA Training Alone Won’t Fix Your Safety Problems.
Accountability Still Has a Place
Whenever this subject comes up, someone assumes that looking at the system means no one can be held accountable. I do not believe that.
There will be cases where an employee knowingly ignores a clear rule. The person may have been properly trained, the correct equipment may have been available, the work may have been reasonably planned, and the expectation may have been consistently enforced. Accountability may be appropriate.
But accountability and root cause analysis are not the same thing.
One asks whether a person’s conduct met the company’s expectations. The other asks what allowed that conduct to result in an injury and what must change to prevent a recurrence.
A company can fairly hold someone accountable and still find weaknesses in planning, supervision, equipment, or safeguards. In fact, it should. Disciplining one employee does not protect the next employee who encounters the same conditions.
Starting with blame also damages the investigation. Employees quickly learn whether the company wants the truth or simply wants a name to place in the report. If they expect every admission to become discipline, they will stop talking honestly about close calls, workarounds, and the way work is really being performed.
That is the same reason an inspection loses its value when people begin changing findings to protect appearances. Honest information is the starting point for improvement. I discussed that issue in When a Safety Inspection Stops Being Honest, It Stops Being Valuable.
Ask What Must Change About the Work
A useful investigation should answer more than who made the last mistake. It should determine whether the safe method was actually possible, whether the correct equipment was present, whether the plan matched the field conditions, whether the same practice had been accepted before, and which safeguard was supposed to prevent the event.
Then the corrective actions should match what the investigation found.
If the problem was knowledge, train and verify competency. If the problem was access, redesign the access. If the problem was equipment, provide the right equipment. If the procedure did not work, revise it and test it in the field. If the shortcut had become normal, management and supervision must address why it was allowed to become normal.
The correction should also be specific enough to verify. “Remind employees to be careful” cannot be measured. “Install a reachable anchorage connector at every designated roof access point and have the superintendent observe the first crew transition” can.
An investigation is not finished when the report is signed. It is finished when the corrective actions are implemented and someone verifies that they work under actual jobsite conditions.
Human error may explain the last action before an incident. It does not explain why one action was able to become an injury.
Training matters. We make our living providing it. But retraining should correct a training problem, not cover up a failure in planning, equipment, supervision, or leadership.
Fix the training gap when it exists. Fix the work when the work is what failed.
