When “Human Error” Is the Root Cause, the Investigation Stopped Early

Tennessee regulators released an eighty-page report this week on the medication error at Ascension Saint Thomas Midtown in Nashville, where four patients scheduled for routine joint replacement were injected in the spine with potassium phosphate instead of an anesthetic. Two are paralyzed. One remains on ECMO. I wrote about the case when it broke and updated that post with what the report found.

The report is said to name human error as the primary cause.

That is true. It is also where the investigation and analysis work starts, not where it ends.

A person did this. A person does almost all of it. Naming the human tells you what happened. It does not tell you why the system allowed it to happen. An investigation that stops there at “human error” has quit at the moment it was about to become useful.

Keep. Asking. Why?

Keep asking

Take the sequence described in the report and keep going.

A technician put potassium phosphate into syringes labeled “mepivacaine.”

Why?

She went to a different storage bin and pulled the wrong vial.

Why was that vial reachable from the workflow that fills spinal syringes? Because high-alert medications were not stored separately.

The hospital announced separate, distinctly marked storage as its fix. Surveyors returned six days later and found the potassium still on the wall with everything else. That's a different chain of whys.

The vial would not scan. Why did that not stop the process? Because the system permitted a manual override without pharmacist verification.

Why did it permit that? Somebody decided, at some point, that an override was necessary. What problem was it meant to solve? Who decided the benefit outweighed the risk? When was that decision last reviewed? Who reviewed the overrides afterward? The report does not appear to identify anyone who did.

She had scanned, entered the lot number, and photographed the medication. Why did none of that catch the problem? Because those steps mostly recorded what the person believed she was holding. Only the scan read the vial itself, and the scan was the one step that got worked around.

She was a float, on a picked-up shift. A technician involved had been given a coaching form that same morning about IV-room workflow and medication calculations. Why was that person pulling medications for spinal syringes? Who was supervising? What is the rule about float staff in the IV room, and does one exist?

Five people touched this defect before it reached a patient, counting the clinician who injected it. Why did none of them catch it?

Because once the wrong drug entered the process, the label, lot entry, photograph, and subsequent checks mostly carried forward the original assumption instead of independently challenging it.

None of those answers is human error. They are questions about the design of the work, the controls around it, and decisions somebody made, usually for reasons that looked reasonable at the time, usually without anyone picturing this particular Friday.

Why human error is such a comfortable place to stop

I do not think surveyors or hospital investigators reach for human error out of laziness. It is comfortable for structural reasons, and they are worth naming, because they operate on your organization too.

It is true. That matters more than it should. A conclusion that is factually correct feels rigorous, and it is hard to argue with someone who says a person made a mistake, because a person did.

It is complete. It closes. Nothing further is implied. You do not have to trace the workflow into another department, examine an old configuration decision, or ask why a known workaround remained available.

The countermeasures are cheap and fast. Retrain. Re-educate. Counsel. Remind. Add a second check. All of those can be finished by Friday, and all of them are documentable, which means they survive the next audit.

The countermeasures are ineffective, though.

And it locates the problem in someone who has often already left the building. If the person who made the error is on leave, the fix can look complete the moment they are gone.

That is the most dangerous version, because the organization experiences relief and mistakes it for improvement.

Dr. Deming put the number at ninety-four percent belonging to the system. You do not have to accept the figure to accept the direction. If your investigations keep landing on the individual, the investigations are the thing to look at.

A test you can run this week

Pull your last ten root cause analyses, or your last ten event reviews, whatever you call them.

Count how many end in retraining, re-education, counseling, competency validation, or a reminder communication. That number is your answer, and in most organizations I have worked with it is high enough to be uncomfortable.

Then count how many changed the conditions of the work. Where a drug is stored. What a system will and will not let a person do. What arrives already prepared so that nobody has to prepare it.

Then look for a single corrective action that names a decision rather than a person.

Not:

“Technician re-educated on scanning policy.”

Something closer to:

“The override permission was granted for a specific operational reason, has not been reviewed since, and now has a named owner responsible for determining whether it should still exist.”

If you cannot find one in ten, the phrase human error is doing work in your organization that it should not be doing.

The sentence at the end

Interviewed a week after the event, the technician told investigators she could not say for certain whether the vial was mepivacaine or potassium phosphate, and that she wished she had looked more closely.

She will carry that sentence for the rest of her life. I have talked to enough people in her position to know that is not a figure of speech.

It may be the most emotionally powerful line in eighty pages.

It is among the least useful if the goal is figuring out how to prevent the next one.

Four patients were harmed. Somebody will write a corrective action plan. The question worth watching is whether it changes where the potassium sits, or whether it changes how carefully somebody promises to look.