Home » Dr. Jayne » Currently Reading:

Curbside Consult with Dr. Jayne 3/28/22

March 28, 2022 Dr. Jayne 7 Comments

image 

The only thing being discussed in the virtual physician lounge this weekend was the trial of Tennessee nurse RaDonda Vaught, who was found guilty of criminally negligent homicide following a medication error. Criminally negligent homicide is a lesser charge than reckless homicide, of which she was found not guilty.

For anyone who hasn’t been following the story, the short version is as follows. Vaught, who was a nurse at Vanderbilt University Medical Center, was managing a patient order for the drug Versed. The patient had been admitted to the hospital’s neurological intensive care unit following a brain bleed. The drug was ordered to help manage anxiety and claustrophobia prior to a PET scan. Vaught couldn’t find Versed in the automated medication dispensing cabinet and used an override to unlock a broader menu of drugs, leading her to obtain the drug vecuronium instead. Versed is the brand name for the sedative midazolam. Vecuronium is the generic name for the drug Norcuron, which is used to aid in “surgical relaxation” for general anesthesia or to create paralysis for patients who are on ventilators in the intensive care unit.

Vaught failed to validate the name of the drug, didn’t notice a warning on the medication vial itself, and didn’t stay with the patient after administration.

Vaught’s attorneys argued that although she admitted making errors with the medication, those errors were part of normal operations at Vanderbilt and reflect systemic dysfunction. Prosecutors alleged that she ignored multiple warning popups, including one that would have said the drug was a “paralyzing agent” and that would have required a reason for the override. Other nurses working on the same unit testified that overriding the medication dispenser was a common occurrence and that a recent EHR upgrade had created delays in obtaining medications from the cabinets. They cited organizational emails instructing nurses to override warnings to reduce delays. Additionally, there was no barcode scanner in the imaging department, where the medication was administered. A scan of the patient’s hospital ID bracelet against the medication might have prevented the fatal drug administration.

As a clinical informaticist and process improvement specialist, I think about these kinds of errors all the time. Our system of having both generic and brand names for drugs causes a lot of confusion. I trained in a residency program where we were only allowed to refer to drugs by their generic names, which probably prevented some errors by newly minted physicians. However, when I entered private practice, there was a lot to learn, as many of my patients referred to their drugs by brand name. I ran across a couple of situations where the patient was on two drugs from the same class that would have been caught had the generic names been used. Fortunately, none of the patients were harmed before we could modify their regimens.

This error also brings up the issue with “look alike” or “sound alike” drugs. In the EHR realm we’ve taken steps to manage the former with interventions such as Tall Man Lettering for drug names ,which help to differentiate names that are close. One could argue that Versed and vecuronium aren’t close other than that they both start with the letter V, but it’s important to understand the level of baseline confusion that might exist when hundreds of drugs are used within a patient care unit on any given day. The practice of medicine has become significantly more complicated over the last two decades with hospitalized patients often being “sicker” than they were in the past. Due to medical advances, patients who previously might have died are living longer, often with a dozen or more drugs to address their health conditions as well as to mitigate issues caused by the drugs themselves.

There are also issues with the setup of the automated drug dispensing cabinet. At the time, the Vanderbilt system only required two letters to be entered to access a drug menu. One organization I worked with had their cabinets set to require five letters to locate a drug. They also had all paralytic agents in a specially colored locked container to make it clear that nurses were accessing something that required additional diligence. Additionally, Vaught was administering medications outside her usual department and didn’t document the administration of the drug. Its lack of inclusion in the medical record led to a death certificate that noted a natural death following a brain bleed rather than being related to the medication administration.

Vanderbilt didn’t report the medication error to the state. Only a year later when an anonymous tip was sent to state agencies did an investigation begin.

Anyone who has practiced in a complex care environment, such as an intensive care unit, understands how even a small distraction can have significant consequences. As a sleep-deprived resident physician, zoning out even for a second could mean missing a critical part of a patient’s information. In past times, ICU nurses may have cared for one or maybe two patients. Today I see them caring for three or more patients, which certainly isn’t helping with attention issues or distraction.

In a hearing before the Tennessee Board of Nursing, Vaught stated that she was distracted while precepting a trainee, but admitted responsibility for the incident. Although this incident occurred long before the COVID-19 pandemic, I can only imagine the level of distraction that nurses faced over the last two years.

In response to this case, I hope all facilities are reevaluating their processes for overrides on automated drug dispensing cabinets and how they store critical drugs such as vecuronium. In speaking with a pharmacist friend about this case, she noted that not all hospitals have increased their search requirements to five characters as my client did. If you’re at one of those institutions, I’d encourage you to quickly prioritize an evaluation of your processes.

This case is a perfect example of the Swiss cheese model of process safety. The more holes that are present, the easier it is for a mistake to happen. When the holes are particularly large, such as when medication overrides have to happen on a daily basis, people become desensitized to the safeguards that are designed to protect patients.

Certainly there were individual actions that led to this tragedy. Policies weren’t followed and literal bright red warnings were ignored. But without the combination of circumstances, the patient would not have received the wrong medication.

Those of us on the clinical front lines have all made mistakes. Some of those mistakes become near misses because of systems that protect patients (and also us as caregivers). But some of those mistakes become true medical errors that have devastating consequences. Comments from my peers run the spectrum from “it’s all Vanderbilt’s fault” to “she deserves the death penalty.” The reality though is that we could all benefit from a closer look, as well as a slower and more thoughtful one, at how a situation like this might unfold in our worlds.

Have you ever been responsible for a medical error? What advice would you give for those who design and maintain the systems upon which you rely? Leave a comment or email me.

Email Dr. Jayne.



HIStalk Featured Sponsors

     

Currently there are "7 comments" on this Article:

  1. As someone who has worked as the main IT support resource for the pharmacy in a hospital and as a technical admin for a clinical application that included a pharmacy module, I would like to call out the role that IT played in the death of this patient. I earnestly believe that if there would have been a barcode scanner on that cabinet and that Vanderbilt required Bar Code Med Amin, BCMA, then this error would not have happened. 100% BCMA compliance would save an unknown number, since med errors are underreported, of patients every year. Clinicians love to shoot IT initiatives with arrows and complain about “tech for tech’s sake”, but this is a pure example of tech making the delivery of care better.

    • Barcode was the first word that popped in my mind. I came out of the VA (non clinical) and I think we were one of the first to use barcoding for meds. I think we pioneered it. I thought this was universal in healthcare now.

    • I agree with all the comments so far.

      And, I also agree that if the process had been followed, the equipment been in place, the software upgrade was not ‘glitchy, the development team had found the issue before deployment and General Availability, she had taken the time to look at the warnings, the hospital administration insisted on fixing the problem instead of ‘advising’ to work around it, on and on and on. That patient would be alive today.

      My point is that these events almost never happen in isolated causalities, there are almost always multiple points of failure, and if I were to look at this from an RCA perspective I would be looking both at the clinical side (Nursing/Pharmacy) and the software development side. The Ishikawa would probably have two sides with about a dozen attributes, ‘fish bones’ on each side, and we could talk extensively about how this failure occurred. With the nurse being just one part of the equation.

      In looking at this, I have never considered administration as part of the causality equation, I think I would add it into future Ishikawa based RCAs.

      On a different but related note about RCAs. You cannot conduct a valid Patient Safety Event RCA if the participants are afraid to speak — if the purpose of an RCA is to identify methods and means to keep the event from happening again, this type of prosecution will do nothing but stifle the improvement of healthcare.

  2. I agree with Jeremy on his statements above, but I will also bring up the fact that even Versed has a monitoring regime.

    I have worked as an RN in Endoscopy back in the old days where a nurse had to be trained on conscious sedation prior to being able to access it. We knew to monitor the patient and what signs required the very specific reversal agent. I may be telling my age, but when I first started my nursing career, we were required to put in 2 years on a med/surg unit before we could ask for specialization. If one wanted to work in the ICU, they received training specific to ICU. Patients on mechanical ventilation were never outside of ICU, but that gradually changed and as it changed the nurses expected to care for them to be trained first. Today, nurses can go straight from boards to ICU without specialized training and are expected to know every brand name and every generic name medication that is in the world today, which is impossible without a smart phone to look it up.

    Now I work with an EHR vendor, and I see so many holes in process I am surprised anyone has survived being a patient at Vanderbilt. Requiring nurses to override every medication in order to get a medication out of the cabinet is a recipe for failure. It causes alert fatigue on top of leadership saying overrides are OK, so why is anyone surprised this happened? What is the policy for the automated Dispensing cabinets? Does the policy respect the current ISMP guidelines for safety? None of this removes her responsibility, and the only reason she was criminally charged is because she is who administered the wrong medication/dose and left the patient unsupervised. Of course, I’d love to know how many nurses think they could give versed and just walk away – and had the Provider that ordered it assumed the patient would be monitored? Why versed for anxiety?

    With all of the above, I am frightened not only for the patients I am also frightened by the guilty verdict in a criminal trial. Nurses are talking about what happened in this case, and everyone should pay attention: Some may not override another medication due to fear of potentially harming a patient and fear of prosecution, others may not report ANY medication error due to fear of prosecution. Will it take skewed override/near miss results to force change? Why are skewed reports more important than a human being.

  3. Criminal prosecution for human error crosses a dangerous line. Johns Hopkins estimated a quarter million medical error deaths per year in 2018. Of those, how many should have resulted in prosecution? In retrospect it is easy to suggest that any given error was negligence so egregious it should be punished. But each of us in medicine make thousands upon thousands of decisions–any one of which might kill. A criminal prosecution here is less likely to improve safety than it is to embitter and alienate caregivers. Good luck improving healthcare with that approach.

  4. My first thought was having two names for every drug has to be confusing. And there has to be a error reporting system without penalty/legal action if you expect to fix the problem. (Denmark).

  5. I’m not happy with the idea of enforcing the use of generic drug names only. Drugs have multiple names and the formularies explicitly acknowledge that.

    My idea is, finding the drug ought to be made easy. Therefore drug searches should support all the drug names.

    The rigor ought to apply to the actual prescription of those drugs. Now, in the online drug ordering workflow, drug searches lead directly to prescription. That reality needs to be addressed one way or another.

    But my gut and my experience suggests to me, having rigid rules on drug searches is the wrong way to go. You are band-aiding the actual problem. And you are doing it by adding to the whole “my EMR is difficult and unfriendly” problem.







Text Ads


RECENT COMMENTS

  1. "Upon learning what I do, several attendees went into some pretty serious rants about how electronic health records have destroyed…

Founding Sponsors


 

Platinum Sponsors


 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gold Sponsors


 

 

 

 

 

 

 

 

 

 

RSS Webinars

  • An error has occurred, which probably means the feed is down. Try again later.