View as “PDF version”
We have done several columns on the tragic incident at Vanderbilt in which a patient died after being inadvertently administered a neuromuscular blocking agent (see our Patient Safety Tips of the Week for December 11, 2018 “Another NMBA Accident”, January 1, 2019 “More on Automated Dispensing Cabinet (ADC) Safety” and February 12, 2019 “From Tragedy to Travesty of Justice”).
The nurse involved in that incident was charged with criminal homicide. That trial has not yet taken place. The Tennessee Board of Nursing recently revoked her nursing license, fined her, and ordered her to pay an inordinate amount to cover prosecution costs (Warick 2021). Perhaps the word “persecution” would be more appropriate than “prosecution”.
While we do not defend the errors and deviations from practice that she made (and she did make multiple ones), when we review a serious incident we always ask “Could another nurse/physician/pharmacist make similar errors, given the set of circumstances this individual acted in?” Our answer to that question in those prior columns was “yes”. Once again, a healthcare worker has been placed at the “sharp end” of a cascade of errors, left with the “smoking gun” and bearing the brunt of blame for a catastrophic event that had multiple root causes and contributing factors.
It's worth reiterating here the findings in our February 12, 2019 Patient Safety Tip of the Week “From Tragedy to Travesty of Justice” and our other previous columns on this incident:
In our December 11, 2018 Patient Safety Tip of the Week “Another NMBA Accident” we discussed a tragic case where a patient was administered a fatal dose of the NMBA (neuromuscular blocking agent) vecuronium instead of the ordered Versed. We discussed a whole cascade of errors, root causes, and contributing factors that were important in leading to the unfortunate outcome.
At the end of that column we noted that the nurse who administered the fatal dose had been terminated from the hospital. While acknowledging that the nurse made several egregious errors, we noted that there were multiple system issues that put her in position to be at the “sharp end” of the error cascade. Now that nurse has been charged with reckless homicide and impaired adult abuse in Tennessee (Kelman 2019).
Clearly, the nurse who was charged made the following errors:
· After entering only the first two letters “VE” for Versed she chose the first drug shown and she overrode a warning in a red box stating that it should be for STAT orders
· She removed the vial from the ADC and did not look at the label which was for vecuronium, not Versed.
· She did not question why she had to reconstitute this preparation (which would not ordinarily have been required if the drug was Versed).
· After administering the medication to the patient in the PET suite, she left to perform another task in the ER, and did not monitor the patient to whom she had administered the medication. (Note that the CMS report does not detail how long she actually observed the patient before leaving the PET suite).
· She failed to document anywhere the details of the drug administration (again, note that there were some system issues that contributed to failure to document).
Though the prosecutors apparently were barred from publicly discussing the merits of the case, a spokesman for the District Attorney’s Office said in a brief statement that overriding the safeguards of the ADC (automated dispensing cabinet) was central to the charge of reckless homicide (Kelman 2019b).
The nurse did override the warning on the ADC screen. But we identified two system issues regarding ADC overrides. First, hospital policy and ADC settings apparently did not require an independent double check for the override.
Second, even more importantly, the warning on the ADC screen was inadequate. The nurse may well have considered this a “STAT” order since she was told to go down to the PET suite now or they will send the patient back here without doing the scan. A proper warning would have said something like “Verify that the patient is intubated and mechanically ventilated or that this drug is being used for an intubation procedure”.
Note that ISMP (ISMP 2019), in its own review of the issue following this case, agrees with us and recommends “Display an interactive warning (e.g., “Patient must be intubated to receive this medication”) on ADC screens that interrupts all attempts to remove a neuromuscular blocker via a patient’s profile or on override. The warning should require the user to enter or select the purpose of the medication removal (“other” should not be a choice) and verify that the patient is (or will be) manually or mechanically ventilated. This type of warning provides an opportunity to specify why the user is being interrupted and requires the user to document a response.”
The CMS inspection report (CMS 2018) also does not mention whether overriding an ADC alert was an issue unique to this nurse or whether ADC overrides had become a routine part of the culture of the unit (i.e. “normalization of deviance”).
We refer you back to our December 11, 2018 Patient Safety Tip of the Week “Another NMBA Accident” for discussion of all the system issues we identified as root causes or contributing factors. But, to summarize, we can identify at least 19 points where interventions or different decisions could have prevented this tragic death:
Ordering the PET scan
We’ don’t know details about the patient’s condition but we’d question whether the PET scan would have been more appropriately ordered as an outpatient, in which case she probably would have been given an oral sedating agent and there would have been no phone call to the Neuro ICU for a nurse to administer an IV sedating agent.
2. A good pre-PET scan checklist could have flagged claustrophobia and the possible need for sedation before the patient was sent to the PET suite. Then, a more informed decision about the need for monitoring might have been made.
Was a PET sedation protocol available?
Recognizing that a substantial number of patients require sedation for PET scans, and that certain sedating agents are contraindicated during PET, there should be formal protocols for PET that take into account how long the sedation is required. Those protocols should include specific drugs and routes of administration as well as guidelines about who needs to be monitored and what monitoring equipment is needed.
Choice of route of administration for a sedating
We wondered why an intravenous agent was chosen rather than an oral agent, particularly since fairly long duration action was desired here. But when we looked at practices of some PET scan units, we found it fairly common that intravenous agents like midazolam were used.
5. A “Ticket to Ride” checklist for intrahospital transport might, likewise, have raised the need for patient monitoring during the procedure.
Decision that patient did not need monitoring
The decision that the patient did not need monitoring was actually made by a different nurse. The PET scan technicians had thought the patient needed monitoring but the patient’s nurse from the Neuro ICU told them she did not need monitoring.
Decision to send a nurse who may not have been
familiar with patient
The nurse who was sent to administer the drug was a “help nurse” who did not have primary responsibility for the patient. It is not clear how much, if anything, she actually knew about that patient. We wonder what sort of “handoff” was actually done between the primary nurse and the nurse sent to the PET suite.
Generic vs. brand name issue
The CMS investigative report indicates the physician order was for Versed (brand name) but the ADC search function defaults to generic names, hence “Versed” did not appear on the ADC search. There should be better consistency between how brand names and generic names are handled in all IT systems. Culture also plays a role here. We do see that, in most settings, healthcare professionals will still call it “Versed” rather than “midazolam”.
ADC did not have an adequate warning
As discussed above, the ADC warning for vecuronium should have been something like “Verify that the patient is intubated and mechanically ventilated or that this drug is being used for an intubation procedure”.
Double checks were not required for ADC
Even though we’ve noted double checks are not without their own problems, an independent double check might well have identified that the wrong drug had been pulled. Note that, ideally, there would have been two double checks (one to override the ADC warning, and one to identify the drug that was removed from the ADC).
Drug vial did not have salient enough warning
The vial in the current case did have a red top and message “WARNING: PARALYZING AGENT”. Yet that was not conspicuous enough to get the attention of the nurse. It seems to us you need to put it in some unique sort of container, perhaps a cardboard or plastic “cage” or something that makes it more difficult to remove as a means of attracting attention to the high-risk nature of the medication.
Nurse failed to look at drug name on the vial
No excuse here. This clearly would have prevented the accident.
Nurse failed to ask herself why she needed to
reconstitute the medication
Simply thinking “I don’t recall ever having to reconstitute Versed before” should have been a clue that this was the wrong drug.
Nurse sent to do 2 separate tasks (ER swallow
test, PET scan sedation)
It almost sounds like the second task was an afterthought “Since you are going down to the ER anyway, why don’t you stop on the way and give this sedation to the patient in the PET suite?”. We’d certainly wonder how the communication actually took place and what details were included in this “handoff”.
Barcoding was not implemented in the
If there had been the opportunity to barcode scan both the patient’s ID bracelet and the medication vial/label, it’s likely that vecuronium would have been recognized as the wrong medication.
Inadequate post-dose observation
The CMS investigative report does not mention how long the nurse stayed with the patient following the administration of what she thought was Versed. Sedation following Versed administration might take a while. Paralysis after injection of vecuronium should be apparent within a few minutes at most. So even observing the patient for a few minutes might have identified the accident. What is not clear is what the patient’s condition was when she was moved from the “injection room” to the “patient room” where she was to stay while the radiotracer was distributed. Did no one notice paralysis?
No physiologic monitoring done
All thought that the medication to be administered was Versed. Injection of Versed or other sedating agent in a patient with an acute neurological condition should have merited monitoring with at least EKG and pulse oximetry. Capnographic monitoring would probably have been difficult in the PET suite. But even though an early rise in pCO2 would have been missed, the patient ultimately would have developed hypoxia and heart rate changes that should have led to alarms and resuscitation.
EMAR or way to document medication
administration in the radiology/PET suite was not available.
Perhaps, if required to do formal documentation of the medication administration at the time of administration, the nurse might have recognized the mistake at a time when rescue of the patient was still possible.
The previous day was a holiday and the PET unit was very busy, likely catching up on scans that might have been done a day earlier. If it were less pressured, perhaps the threat to send the patient back without a scan (if she could not be sedated immediately) might not have taken place.
It is conceivable that a different action or decision at any one of the above 19 points might have prevented this tragic accident. Yes, several of these relate directly and solely to the nurse. But the majority are enabling factors that can allow a human error to break through multiple defense systems and lead to patient harm.
In our December 11, 2018 Patient Safety Tip of the Week “Another NMBA Accident” we concluded that it was quite conceivable that another nurse might have made similar errors given the same set of circumstances and contributing factors. That is, the many system issues actually put that nurse in a position where human error would leave her at the “sharp end” of the error cascade. As such, we would not have recommended terminating the nurse. And we think that charging this individual with reckless homicide is an outrageous travesty of justice. This nurse will undoubtedly live with the memory of this terrible accident the rest of her life. But to put her through the rigors of a homicide trial, even if she is ultimately acquitted, is simply wrong.
And the current revocation of here nursing license, especially in view of the upcoming criminal trial, is particularly bothersome.
ISMP (Institute for Safe Medication Practices) recently expressed a stance similar to ours (ISMP 2021). They also stressed a few important considerations. First, they note there was undoubtedly significant outcome bias in the Board’s decision, noting that “the Board has not filed disciplinary action against all TN nurses who have not read a medication label carefully, obtained a nonurgent medication from an ADC via override, drawn an incorrect conclusion, failed to monitor a sedated patient, or failed to document a medication error in the patient’s record.” In this case, the patient died, so the decision was likely biased because of that outcome.
A second important consideration deals with the concept of Just Culture. ISMP notes that the Board apparently has a policy that differentiates between human error, at-risk behavior, reckless behavior, and bad intent. Apparently, the prosecutor noted that the nurse did not act with bad intent but alleged that she did act recklessly. ISMP argues that her actions were either unintentional (human error) or at-risk behaviors, not reckless behaviors. ISMP notes that conscious disregard of a substantial and unjustifiable risk is a requirement for reckless behavior, and notes that this nurse had no idea that she had made a mistake. Failure to read the front of the medication label was either due to a momentary distraction (error) or an unsafe practice habit (at-risk behavior). ISMP argues that the Board did not determine whether the nurse saw the risk(s). They also note that, while she made a conscious decision to not monitor the patient or scan the medication’s barcode, she was told that monitoring was not required, and barcode scanning technology was not available in the radiology suite.
ISMP further argues that the Board relied on an incomplete investigation of the event, particularly related to the question, “What normally happens in similar circumstances?” ISMP would have wanted to know what happens “normally” for patients who were anxious about radiology scans due to claustrophobia. It would have asked whether they were given oral anxiolytics or IV sedatives, and whether they were monitored and by whom and for how long? ISMP also notes the Board failed to consider all the system deficiencies that they have outlined, most of which we have discussed in our prior columns.
ISMP also notes this nurse was contrite and honest in admitting errors she made and was devastated about the tragic patient outcome.
The ISMP column has a useful table outlining the nurse’s alleged failures, with comments about each specific “charge”, noting that several were “disputed” failures and others “undisputed failures” but adds comments as to whether the latter were due to reckless behavior or due to human error or at-risk behavior in accord with Just Culture principles. It is worth your taking time to read the full ISMP column and our prior columns listed below.
Every day in every hospital, healthcare workers make some of the mistakes or deviations noted in this incident. Fortunately, in most cases, harm does not come to patients. But when systems fail, adding contributing factors and lacking safety defenses, bad outcomes ensue. Did the Tennessee Board of Nursing, the hospital, and the prosecutors in the upcoming trial really consider that fundamental question “Could another nurse/physician/pharmacist make similar errors, given the set of circumstances this individual acted in?” We think anyone considering the totality of circumstances in this case would agree with both our view and ISMP’s view that the actions against this nurse (now by multiple parties) have been unjust.
Some of our prior columns on neuromuscular blocking agents (NMBA’s):
June 19, 2007 “Unintended Consequences of Technological Solutions”
July 31, 2007 “Dangers of Neuromuscular Blocking Agents”
November 2007 “FMEA Related to Neuromuscular Blocking Agents”
January 31, 2012 “Medication Safety in the OR”
February 7, 2012 “Another Neuromuscular Blocking Agent Incident”
October 22, 2013 “How Safe Is Your Radiology Suite?”
December 9, 2014 “More Trouble with NMBA’s”
December 11, 2018 “Another NMBA Accident”
January 1, 2019 “More on Automated Dispensing Cabinet (ADC) Safety”
February 12, 2019 “From Tragedy to Travesty of Justice”
April 2019 “ISMP on Designing Effective Warnings”
February 2021 “ISMP: 2 Alerts on NMBA’s”
Warick S. Former nurse accused of giving fatal dose of wrong medicine to patient loses nursing license. News 4 Nashville 2021; Jul 26, 2021
Kelman B. Vanderbilt ex-nurse indicted on reckless homicide charge after deadly medication swap. Nashville Tennessean 2019; Published February 4, 2019
Kelman B. Vanderbilt nurse: Safeguards were ‘overridden’ in medication error, prosecutors say. Nashville Tennessean 2019; Published February 7, 2019
ISMP (Institute for Safe Medication Practices). Safety Enhancements Every Hospital Must Consider in Wake of Another Tragic Neuromuscular Blocker Event. ISMP Medication Safety Alert! Acute Care Edition 2019; January 17, 2019
CMS (Centers for Medicare and Medicaid Services). Statement of Deficiences. Complaint #TN00045852. CMS 2018; Date of survey 11/08/2018
ISMP (Institute for Safe Medication Practices). TN Board of Nursing’s Unjust Decision to Revoke Nurse’s License: Travesty on Top of Tragedy! ISMP Medication Safety Alert! Acute Care Edition 2021; 26(16): August 12, 2021
Print “PDF version”