View as “PDF version”
The tragic death of a prominent cancer researcher serves as a stark reminder of things that can go wrong with infusion pumps (Evans 2020). The 61-year-old man was admitted to the hospital with a stroke. Over the next two days his condition stabilized and he began speech, physical and occupational therapies. His diet was advanced to solid foods. But, as plans were being made for his move to a rehab facility, his condition took a turn for the worse and he failed a swallowing test. His breathing became labored and a feeding tube was inserted due to his difficulty swallowing. During the procedure, he apparently aspirated a large piece of food from his stomach. It lodged in his airway and he suffered acute respiratory arrest and was placed on a ventilator.
To sedate him while he was on the ventilator, he was given four small doses of fentanyl, ranging from 25 to 75 micrograms. At that time he was responsive and able to follow simple commands. It was anticipated he would come off the ventilator soon.
A little after 11 PM, a nurse covering for his primary nurse responded to an alarm going off on a medication pump next to his bed. An IV bag of Lactated Ringer's solution needed to be replaced. The nurse obtained a new bag of the solution and hung it on the pole along with other medications also attached to the pump. One of those bags contained the fentanyl, which had been discontinued 12 hours earlier. The nurse scanned the bag of Lactated Ringer's solution and a barcode that came up on the pump before restarting it at 11:24 p.m. She had attached the Lactated Ringer's solution to one channel on the pump, but mistakenly pushed the start button for another channel — the one attached to a bag of fentanyl. As soon as she pushed the start button on the module that was actually going to give the fentanyl, the alarm went off. But, instead of checking the alarm and reading the alarm code, the nurse just hit a silence button. A few minutes later, another nurse who checked the pump also missed the error.
Because the mix-up went unnoticed, over the next 69 minutes he received 1119.4 mcg of fentanyl. The error was finally discovered at 12:33 AM. Narcan was not given then or when a doctor came to his bedside around 2 AM, when he was profoundly hypotensive. He never regained consciousness and was subsequently removed from life support. The family apparently was not made aware of the mistakes until details were revealed during a medical record review done as part of a malpractice investigation.
We, of course, do not know details about the hospital’s own root cause analysis (RCA) or incident investigation but it should have focused on at least the following issues:
Our March 2020 What's New in the Patient Safety World column “ISMP Smart Infusion Pump Guidelines” discussed ISMP’s “Guidelines for Optimizing Safe Implementation and Use of Smart Infusion Pumps” (ISMP 2020). The hospital violated two of the most important recommendations in that guideline.
One point that was emphasized in those guidelines was performance of double checks at certain points. When starting selected facility-defined high-alert medication infusions and at additional facility-defined steps (e.g., change of shift/handoffs, change in the rate/dose of infusion, change in bag/bottle/syringe) hospitals should require that a double check be performed and documented to verify the following before starting the infusion:
2. Patient weight used for weight-based medications
4. Drug concentration
5. dose-rate of continuous infusions, dose and rate of intermittent infusions
6. Channel selection
7. Line attachment
ISMP noted that technology (eg. barcoding) is the preferred method for double checking most of these items but use of a second practitioner is necessary to perform and document an independent double check for line attachment. (Keep in mind that we also always recommend that, when multiple lines are connected to multiple different sources, it is critical that good labeling be placed not only on the sources but also on the lines themselves).
It also emphasized another often overlooked point: “Immediately discard all continuous IV medications and epidural infusions (e.g., continuous infusions of magnesium, neuromuscular blocking agents, and opioids) after discontinuation (i.e., do not leave hanging on an IV pole or at the bedside).”
While those 2 issues provide, arguably, the most important lessons learned from this tragic incident, keep in mind that the cascade of errors in this case is typical of most adverse events where there is a bad patient outcome. Almost each of the issues above represented an opportunity to intervene and stop the cascade before an unfortunate outcome occurred.
In our March 5, 2019 Patient Safety Tip of the Week “Infusion Pump Problems” we did an extensive review of safety issues related to infusion pumps. We hope you’ll go back to that column to review our many recommendations.
There was also another recent review of human-based errors involving smart infusion pumps (Kirkendal 2020). The authors classified five main categories: undocumented orders, drug library errors, programming errors, administration errors, and ancillary equipment errors. Within these five main categories of errors, the investigative team identified specific types and even more granular subtypes.
While they discussed issues like programming errors and drug library errors, they also stressed several other important issues: misunderstanding or misinterpreting pump alerts, forgetting to close tube clamps, and bag misalignment with multiple infusions. And they acknowledged that some errors may be the result of workarounds performed by clinicians to intentionally bypass safety features.
Lastly, we are strong believers in transparency and disclosure and apology Not only is that the right thing to do, but one can only wonder if prompt disclosure and apology might have prevented the details of this tragic case from being splashed in the headlines. Our many columns and resources on disclosure and apology and communication and resolution programs are also listed below.
Our prior columns related to infusion pump issues:
· May 2010 “FDA's Infusion Pump Safety Initiative”
· April 27, 2010 “Infusion Pump Safety”
· August 2016 “Home Infusion Therapy Pitfalls”
· March 2020 “ISMP Smart Infusion Pump Guidelines”
· August 4, 2020 “Intravenous Issues”
Some of our prior columns on Disclosure & Apology:
July 24, 2007 “Serious Incident Response Checklist”
June 16, 2009 “Disclosing Errors That Affect Multiple Patients”
June 22, 2010 “Disclosure and Apology: How to Do It”
September 2010 “Followup to Our Disclosure and Apology Tip of the Week”
November 2010 “IHI: Respectful Management of Serious Clinical Adverse Events”
April 2012 “Error Disclosure by Surgeons”
June 2012 “Oregon Adverse Event Disclosure Guide”
December 17, 2013 “The Second Victim”
July 14, 2015 “NPSF’s RCA2 Guidelines”
June 2016 “Disclosure and Apology: The CANDOR Toolkit”
August 9, 2016 “More on the Second Victim”
January 3, 2017 “What’s Happening to “I’m Sorry”?”
October 2017 “More Support for Disclosure and Apology”
August 13, 2019 “Betsy Lehman Center Report on Medical Error”
September 2019 “Leapfrog’s Never Events Policy”
Other very valuable resources on disclosure and apology:
Evans T. Drug error at Eskenazi Hospital killed prominent cancer researcher. Here's how it happened. Indianapolis Star 2020; October 30, 2020
ISMP (Institute for Safe Medication Practices). Guidelines for Optimizing Safe Implementation and Use of Smart Infusion Pumps. ISMP 2020; February 10, 2020
Kirkendal, ES., Timmons K., Huth H, et al. Human-Based Errors Involving Smart Infusion Pumps: A Catalog of Error Types and Prevention Strategies. Drug Saf 2020;
Print “PDF version”