Friday, December 7, 2018

A fatal medication error

A 75-year-old woman died at Vanderbilt University Medical Center after receiving intravenous vecuronium, a drug that causes muscle paralysis, instead of Versed, a sedative.

Here’s what happened.

She had been recovering well from an intraparenchymal brain hematoma after a fall. While awaiting a full body positron emission tomography (PET) scan in the radiology department, the patient said she was anxious about being in the machine because she was claustrophobic.

A doctor ordered Versed 2 mg IV in the electronic medical record at 2:47 PM. Two minutes later, the pharmacy verified the order. The radiology department staff said they could not give the medication because they were very busy and the patient would need to be monitored.

So nurse A, who was functioning as the “help all” nurse was asked by the patient’s nurse, nurse B, to go to radiology and give the medication. At 2:59 PM, nurse A went to the automated dispensing cabinet (ADC) and searched for Versed in the patient’s profile. When she couldn’t find the drug listed, she selected the “override” setting to search for the drug. She entered the first two letters of the drug, VE, and clicked on the first medication that popped up.

She took out the vial and read the instructions on how to reconstitute the drug but did not recheck the name of the drug. She put the vial and the other items needed for the injection into a baggie, added a sticker with the patient’s information on it, and wrote “PET scan, Versed 1-2 mg” on it.

Nurse A then went from the ICU to the radiology department. During an interview with Centers for Medicare and Medicaid Services (CMS) investigators, she mentioned she had never been to the PET scan area before. She identified the patient, reconstituted the drug, gave it, but did not monitor the patient after the injection. She went back to the ICU and gave nurse B the baggie containing the remaining drug.

A few minutes later, a transporter noticed the patient was unresponsive. The rapid response team was called.

The two nurses heard the call, went to radiology and saw the patient had been intubated and resuscitated. Nurse B then noticed that the baggie contained vecuronium instead of Versed and asked nurse A if that was what she had given the patient. Nurse A confirmed vecuronium was given and admitted her mistake. She did not enter the vecuronium injection in the medical record because she was told the system would automatically do it. [It did not.]

The patient suffered severe anoxic brain damage and died the next day.

Nurse A was fired a few days after the incident.

Key points

Nurse A apparently did not see that “WARNING: PARALYZING AGENT” was printed on the red top of the vecuronium vial found in the baggie.

The actual amount of vecuronium the patient received was unknown because nurse A could not recall how much she had drawn up or injected.

Nurse A did not adhere to the five rights of medication administration—right patient, right drug, right dose, right route, and right time.

Nurse A may not have been familiar with Versed which, unlike vecuronium, does not need to be reconstituted.

The hospital’s ADC was searchable using either a generic or brand name, but not both. The machine’s default search is by generic name.

Even if the patient had been given Versed, she would have required monitoring afterward. The hospital did not have a policy stating how patients should be monitored after receiving sedation or how often they should be monitored.

When the death was reported to the medical examiner, the physician who made the call said the medication error was “hearsay” because it was not documented in the medical record, and the medical examiner declined the case. According to the death certificate, the cause of death was the cerebral bleed.

CMS had considered discontinuing Medicare payments to Vanderbilt because of the incident but after its investigation, accepted the hospital’s plan of correction.

What do you think—system error, human error, or both?

Sources:

CMS investigation report
The Tennessean

22 comments:

Huw said...

I am based in the UK where it's very rare to use drug brand names and almost all hospital prescribing is generic. I had to look up Versed (it's midazolam). I would imagine this culture increases the risk of errors significantly as the possible permutations of drug confusion magnifies exponentially.

Skeptical Scalpel said...

Unknown, I agree with you. As this case illustrates, our system can lead to mix-ups especially when the ADC does not permit a simultaneous search for both names.

Debra Gottsleben said...

Sounds like both to me. But think Nurse A didn't do some very basic checking. She was in an area unfamiliar to her so I would expect her to be extra careful. The entire incident is very scary.

Anonymous said...

The most alarming finding is the criminal cover-up and the forgery of the charts and death certificate. Alarming, but unfortunately, not surprising at all, nor does the lack of criminal invesigation.

I wonder if those Norwegian auditors would have ruled this death as unavoidable according to the charts?

Glen said...

Given how common claustrophobia is, Radiology should have staff trained to administer and monitor sedation. Reducing motion artifacts would reduce re-exams and reduce exposure.
ALARA/Radiation safety procedures and MRI safety zones are confusing for outsiders. A person already trained and familiar with the rules could administer medication and monitor patient with much less distraction.
If sedation order is included as part of radiology order in EMR, the radiology department should have medication and person ready when patient arrives.

artiger said...

Human and system error, to be sure, but human>system in this one. But, as in so many things, these days, it's not the mistake, but the cover up that kills. CMS appears to be a bit more forgiving.

Skeptical Scalpel said...

I agree nurse A did not check things. As I said in the post, where were the five rights?

I’m not sure what the Norwegians would have done with this. Does it happen frequently? Such cases are obviously hard to quantify.

The radiology staff staff said they couldn’t monitor the patient because they were too busy. Unfortunately nurse A didn’t monitor her either.

No question the cover up made it worse.

Clark Venable said...

The event occurred on 12/26/17. The CMS report, page 32, quotes a pharmacist as saying:
"We rolled out EPIC, our new system for documentation last year in November [2017]."

And nurse # 1 herself is quoted as saying:
" I spoke with [Named Nurse Manager] and he/she told me the new system would capture it on the MAR [Medication Administration Record]."

So, it sounds like Epic had been live for a month, maybe two. No, I'm not blaming Epic, but it was still a 'new' system and RN #1 had a student (hence divided attention).

Skeptical Scalpel said...

Clark, thanks for pointing that out. I missed it. I'm not sure how much the ADC and Epic interface was a factor but certainly it could have been.

Oldfoolrn said...

Sounds to me like something that occurred in a back of the woods dispensary not an academic medical center where resuscitation should be immediately available.

Skeptical Scalpel said...

Old, I think it took them a while to realize she had stopped breathing. No matter how good the resuscitation was, it probably started too late.

Anonymous said...

The report said, "the drug appears to have caused the patient, who was otherwise stable, to lose consciousness, suffer cardiac arrest and ultimately be left brain dead"

Unfortunately, the patient would have been conscious with air hunger for some time before hypoxia caused her to lose consciousness.

Skeptical Scalpel said...

Anon, I agree with you that this must have been a terrifying death. To be paralyzed but completely awake is a nightmare.

Joe Kersey said...

As a retired anesthesiologist who avoided a lot of the computer generated records and drug dispensing (good grief -- it sounds like a Coke machine in an airport lounge -- and who used a lot of midazolam (always referred to as such rather than Versed) snd some vecuronium (although I preferred pancuronium and -- yes -- curare; I trained with those after all); there was a huge power fight about who was allowed to use neuromuscular blocking agents in an ER or ICU setting for a few months in our hospital. Clearly the victory lay on side of logistically it's just easier to let more folks use those drugs. I hope that the nurse "diluted" the vecuronium such that she got the full 10 mg of the vial right off the bat rather than struggling with 2 mg after all the antecedent medicines she had had in her hospital story which would have let her suffer like a beached fish before dying of exhaustion.

Skeptical Scalpel said...

Unknown, sorry for taking so long to respond. I remember curare. Because of the need for paralysis in emergencies, it must be available in units on code carts. Since it’s unclear how much vec she got, we’ll never know if she suffered for a long time before she stopped breathing.

Skeptical Scalpel said...

“It” meaning vecuronium in my comment above not curare.

Debra S. Wright said...

Human error...no question.

Skeptical Scalpel said...

Ultimately, yes.

Anonymous said...

I believe that it was both a human and system error. On a side note, I have a question. So, I am not an anesthesiologist, but, wouldn't 4 mg need to be given of Vecuronium to causing paralysis in a patient???

From my calculations, 1 mL of 1mg/mL Vecuronium is a low dose of this medication that should be given at .08 to 0.1mg/kg. If the medication was reconstituted to the usual dose of concentration of 1mg/mL and the client was given 1 mL, she would have received 0.020042 mg/kg (less than a paralytic dose of the client weighing 110 lbs). Assuming that the deceased client weighed 110 lbs (a low estimated weight), she would have had to receive 4mL of the 1mg/mL concentration to have a paralyzing dose of 0.08mg/kg. Therefore, it is unlikely that this medication actually was the cause of this client’s death. Unless, the nurse got her NS flush syringe and medication syringes confused and gave the entire 10 mg of Vec...

Skeptical Scalpel said...

Anon, thanks for commenting. As I said in the post, the nurse wasn’t sure how much she gave. The patient expired and no other explanation for the death has been offered. An autopsy was not done.

Unknown said...

I still would like to know if my above statement is possibly correct? I am having a hard time finding information on sub-paralytic doses for Vec., other than vocal chord damage.

As state above, if the medication was reconstituted to the usual dose of concentration of 1mg/mL and the client was given 1 mL, she would have received 0.020042 mg/kg (less than a paralytic dose of the client weighing 110 lbs). Assuming that the deceased client weighed 110 lbs (a low estimated weight), she would have had to receive 4mL of the 1mg/mL concentration to have a paralyzing dose of 0.08mg/kg. Even if RN# 2 gave 2mg that would be 0.040084mg/kg. Therefore, it is unlikely that this medication actually was the cause of this client’s death. Unless, the nurse got her NS flush syringe and medication syringes confused and gave the entire 10 mg of Vec. Below, explains why I am asking this question.

On page 10 of the CMS report it states:

RN #1 was asked how much medication did
he/she administer to Patient #1, and the RN
stated, "I can't remember, I am pretty sure I gave
[him/her] 1 milliliter.
RN #1 was asked what was done with any left
over medication, and the RN stated, "I put the left
over in the baggie and gave it to [Named RN #2]..."

Then, on page 35-36 of the CMS report:

Observations in conference room 167 on 11/2/18
at 12:35 PM revealed one (1) clear zip lock
baggie with an orange biohazard label. There was
handwriting on the baggie in a pink color marker
that documented, "Versed 1mg 2mg PET
1251." Inside the baggie was a vial with a few
drops of clear liquid remaining in the vial. The vial
was labeled Vecuronium Bromide 10mg. 1mg/mL
when reconstituted to 10mL... There was one (1) 10 mL syringe
labeled "Normal Saline" with a capped needle
attached. The syringe had 8 mL of a clear liquid
remaining in it. There was one 10 mL syringe labeled "Normal Saline" with 1.5 ml of a clear
liquid remaining in it and capped with a white cap
with no needle. There was also a 2" alcohol prep
pad in the baggie...

Furthermore, on page 51 of the CMS report indicates that Vecuronium was given during RSI and ACLS before returning to the ICU by stating:

A physician progress note written on 12/26/17 at
6:28 PM by an Advance Practice Registered
Nurse (APRN) and co-signed by Physician #2
documented, "...Received patient to NCU [Neuro
Critical Care Unit] after cardiac arrest in PET
scan. Per report, ROSC [Return of Spontaneous
Circulation] received after approximately 2 rounds
of ACLS [Advanced Cardiac Life Support]. Patient
was intubated during event...Current
Facility-Administered
Medications...Vecuronium..." No dose, route or
frequency was documented.

Therefore, the autopsy would have show that Vecuronium was given but might not have been able to determine how much RN #1 gave because no one ever documented the the amount of Vecuronium that, including the code team, except in the progress note by the APRN.

The main problem as I see it was the lack of a medication scanning system in radiology, as well as the new rollout of epic EMR. IMO, the hospital is throwing this nurse under the bus because she was the only one who admitted she made a mistake and was experiencing extreme emotional distress, feeling of guilt, and shame. Moreover, I doubt that there was a red dust cover on their Vec vials because, as a former CVICU & CTICU RN and an NP student, I never keep the dust covers of my used vials. The whole report reads like a massive cover-up. The hospital threw this nurse under the bus because CMS came and threatened to take away all their funding.

See the full CMS report here: https://www.documentcloud.org/documents/5346023-CMS-Report.html#document/p9

Skeptical Scalpel said...

Sorry for the delay in responding, I think your calculation is correct, but again is unclear how much of the solution the patient actually received. Also, she suffered a cardiac arrest. It would have been quite a coincidence if it wasn't related to the injection.

As I said in the post, somewhat overlooked in the discussion is the fact that even if she had received Versed, she should have been monitored.

Post a Comment

Note: Only a member of this blog may post a comment.