A US television commercial for Berlitz language courses begins with a senior officer briefing a draftee on how to man the station before leaving him on his own. No sooner had the officer left than a desperate call crackled over the radio: ‘May day! May day! Hello! Can you hear us? We are sinking. WE ARE SINKING.’
The novice slowly and deliberately replies in his second language, ‘Hello, zis iz za German coast guard.’
With no time to spare, the party in peril fires back, ‘We are sinking. We are sinking!’ To which the young radio pal replies, ‘What are you zinking about?’
Admittedly, I speak only English and many (perhaps most) reading this article have English as their second language. Please bear with me as I use my language to share a few things we are learning in the US about improving patient safety.
I’ve been thinking about how easy it is to get things wrong.
At my grocery store, the 1 per cent and non-fat milk containers sit next to each other in the cooler. They look alike except that one label is a lighter shade of blue than the other. Don’t ask me which one. I can’t remember. More than once, I’ve arrived home with the wrong product.
The same is true for medications in hospital pharmacies. Not only the containers but the names also often look and sound alike.
According to a 2006 survey by the American Society of Health-System Pharmacists (ASHP), the number-one fear gripping patients (61 per cent) entering hospitals is that they will be given the wrong medication. Their fears are reasonable. According to the Institute of Medicine, one in five medication administrations involves an error, resulting in as many as 1.5 million adverse drug events each year.
An estimated seven thousand of these errors result in death, roughly the equivalent of 6.5 plunging Titanics, or an AirBus 330 crashing every three weeks. If planes and ships went down at this rate, we would all stay home.
For over a decade, The Joint Commission’s (TJC) yearly National Patient Safety Goals have had the same number-one goal: ‘Improve the accuracy of patient identification,, requiring hospitals to use ‘at least two patient identifiers when providing care, treatment, or services.’ TJC’s third perennial patient-safety goal: ‘Improve the safety of using medications,’ requiring hospitals to annually review a list of look-alike/sound-alike medications and to ‘take action to prevent the interchange of these medications.’ This short list illustrates their point:
- Cerebyx vs. Celebrex
- Cedax vs. Cdex
- Dioval vs. Diovan
- Diovan vs. Darvon
- Dobutamine vs. Dopamine
- Epinephrine vs. Ephedrine
Can you read the paragraph below?
Apaprenlty, the oredr of the ltteers in a wrod deosn’t raelly matetr that mcuh. The olny iprmoatnt tihng is taht the frist and ls a tltteer be in the rghit pclae. The human mnid deos not raed ervey lteter by istlef, but the wrod as a wlohe. Evreywrod in tihs paragrahp wtih fuor or mroe letetrs is missepled.
For all those whose first language is English, the paragraph above was an easy read. This suggests that the more familiar a person is with drug names, the greater the possibility of getting confused with them. Did you notice with the drug names above how not just the first and last letters were the same but also the first and last syllables were similar? As a lay person, the drug names on the above list do not look alike to me, nearly as much as they do for pharmacists or nurses who know them well.
The same confusion happens with patient names. Not only do some names look and sound alike (e.g., Smyth, Smith); some patients in the same hospital may have similar or same names. Nearly five million people in the US have the surname Smith or Jones. In China, around 180 million people are surnamed Wang or Li.
Even with rare names, mix-ups occur. On his way to surgery in a 200-bed hospital, my uncle had a medical chart on this chest bearing his name. My aunt complained that someone used the wrong middle initial. ‘It’s N,’ she protested, ‘not A.’ Turns out that David N Neuenschwander was heading to surgery accompanied by a chart belonging to one David A Neuenschwander.
No patient wants to be affected by a medical error, nor does a caregiver.
During a typical business day, nearly six billion bar codes are scanned around the world with virtually no errors. Prior to point-of-sale bar-code scanning, IBM studies involving the review of cash-register receipts from customers leaving supermarkets showed that one in ten entries had an error.
Scanning at the point-of-sale was born in the United States forty years ago on June 24, 1974. The first bar code read by a scanner on a retail product was a pack of Wrigley’s Juicy Fruit chewing gum. Within ten years, virtually all items in grocery and drug stores had bar codes. Soon, dresses at department stores and wrenches at hardware companies had bar codes. Scanning was being used to track packages, identify cattle and luggage, and admit fans to concerts and football matches. Bar codes were everywhere on everything - except on patients and drugs in hospitals.
Two decades after the Juicy-Fruit scan, a nurse named Sue Kinnick was vacationing in Seattle. Upon returning a Hertz car at the airport, she was impressed that the agent could retrieve her rental record just by scanning the bar code on the windshield. In fact, she was so impressed that she nearly missed her flight to Kansas by asking questions. Upon returning to work, she asked the medical centre’s IT department if they could create something like this for positively identifying patients. In 1995, bar-code scanning at the point-of-care (BPOC) was born. To make it work, the hospital had to apply its own bar codes to all drug packages, an ominous and high-risk practice.
Bar codes did not appear on hospital drugs until 1991 and then only on a dozen products. It would take another fifteen years before all drugs would have bar codes but not without first convincing the US Food and Drug Administration (FDA) to issue a rule requiring drug manufacturers to include them on all immediate packages effective April 2004. This proved to be the tipping point for the widespread adoption of scanning at the point-of-care.
Safe practice requires comprehensive bar coding at the source, which rarely happens voluntarily. In most cases, source labelling requires government regulations.
Today nearly two-thirds of US hospitals are scanning patients and most medications and the results are impressive.
A landmark study from the Brigham and Women’s Hospital in Boston, published in the New England Journal of Medicine, demonstrated the benefits that other ‘bar-coding’ hospitals had claimed to experience:
- 41 per cent reduction in non-time-related medication errors
- 51 per cent decrease in non-time-related potential adverse drug events
- 27 per cent decrease in time-related medication errors
Scanning medications is but the nose of the camel that we need to shove into the bar-code tent. Expect to see more and more hospitals implement scanning at points of collection to properly label and track blood, stem cells, specimens, biopsies, x-rays, and mother’s milk. Scanning will be increasingly utilised for matching patients with transfusion, implant, meal, TPN, medical device, and other medical service requirements, as well as for matching babies with mothers and ensuring that caregivers have the right documents for the right patients. The list is endless.
The next gap we are addressing in US hospitals is upstream in the medication-preparation process. While most medications can move from the loading dock through the pharmacy to the point of care intact with manufacturer bar-coded labels, some medications require preparation and compounding. These include the higher-risk drugs like chemotherapy, often involving high-risk patients (e.g. paediatrics, oncology, etc.)
In my lectures, I show a quick video clip of my topping off a paper cup with Dr. Pepper at a self-service soda fountain. Then I ask the audience what was in my cup. Not everyone pays enough attention to get the answer right. So I show the clip again. Knowing the question, they pay close attention, and everyone gets it right. Actually, everyone sees that they were wrong when I show them the full video, in which I randomly add ten sodas into the cup before I reach Dr. Pepper. In the US, we call this drink a ‘suicide.’
Then I show an IV bag that has been prepared in a pharmacy and ask, ‘What’s in the bag?’ They don’t know, of course, and neither do nurses. They must trust that their pharmacies have followed the orders.
BPOC cannot determine if bags have been compounded and labelled correctly. And the fact is, a significant portion is not. I’ve seen studies falling under the 5 to 10 per cent error range.
What if FedEx were to have as high an error rate? They would not be in business. But even when FedEx gets the right-labeled boxes to the right customers, what if Amazon has filled 10 per cent of these boxes with wrong items in the warehouse? Amazon would be out of business.
The fact is when Amazon fulfils orders, they have an error rate of .001 per cent, which is achieved with bar-code order verification systems.
Scanning at the point-of-care cannot fix compounding errors, but scanning ingredients in the pharmacy can prevent compounding errors.
Though presently fewer than 5 per cent of US hospitals are using bar-code enabled medication preparation (BCMP) systems to compound sterile products, I expect most will do so by the nd of this decade.
A colleague and I have just completed an extensive report on BCMP technologies available in the US, reviewing over a dozen products, including semi-automated manual and highly robotic systems. Both types use bar-code scanning to verify ingredients and to produce final order-specific bar-code labels for scanning prior to administrations.
Semi-automated manual systems utilise a combination of imaging and gravimetrics. Cameras capture images of drug vials and drawn syringes during compounding, which pharmacists may view when checking completed orders. Some systems utilise gravimetric verification tools, which require compounding technicians not only to scan products to ensure they have the right drugs and diluents but also to weigh vials, syringes, and bags before and after draws and injections. These steps verify that the right amounts of ingredients have been used. This is analogous to self-checkout grocery stands. One-by-one, shoppers scan products and place them in bagging stations. If the weights don’t match the items scanned, shoppers cannot proceed.
With highly automated robotic systems, technicians must manually place drugs, diluents, and containers onto staging platforms from which robotic arms retrieve items and complete the compounding process.
One leading BCMP provider compiled data from several million IV preparations, which revealed a consistent product selection error rate of nearly 5.5 per cent distributed across 14 different error types, the most repeated of which involved wrong selection of products. While these data are frightening, the fact that BCMP technology intercepted the errors in pharmacies before reaching patients is encouraging.
For years, I have argued that bar coding at the point-of-care is to patient safety is like what seatbelts are to passenger safety—not the only thing but a salient thing. Cars must have good brakes, clean windshields and drivers should be sober, and drive defensively. Caregivers must be well rested, highly trained, undistracted, and attentive. But when everything is said and done, accidents happen and bar coding can keep caregivers and patients from flying through windshields.
I will not rest until we view scanning the patients in hospitals as important as buckling up before putting our cars in gear. I will press on until we view failure to scan infants before administering heparin as horrifying as driving off without having secured them in safety seats.
A 2012 study by the Automobile Association of America revealed that three of four infant safety seats are improperly installed in automobiles. Many parents, committed to the welfare of their children, have the illusion that their little ones are safe. Kids may be properly placed in car seats, but if those seats have not been installed correctly, the seats and the children are at risk of going through windshields.
Taking things full circle, scanning at the point of care may likewise have a false sense of safety if the hospitals are not utilising bar-code safety systems when compounding medications in their pharmacies.
As hospitals attempt to make their ships safer for patients, I hope they will also consider the benefits to other invaluable passengers on board.
In 2010, here in Seattle, a loving nurse, with 25 years of exemplary service, accidentally overdosed an infant with sodium chloride. Six months later, unable to endure the grief, she took her own life. Four years earlier, on her second birthday, Emily Gerry’s life was lost as a result of a compounding error in the hospital’s pharmacy. The pharmacist on duty was charged with a felony, spent time in jail, and lost his license to practice his life calling and passion. Such mistakes could have happened to any nurse or pharmacist. Both could have been prevented by using proven bar-code systems, one at the point of care, the other upstream during the preparation process. These deaths were unnecessary.
We owe everyone on board the safest journey possible—patients and caregivers alike.
That’s what I’ve been thinking. What do you think?
- To Err Is Human, Institute of Medicine, Nov 1, 1999.
- “In the Clean Room: A Review of Technology-Assisted Sterile Compounding Systems in the US,” Mark Neuenschwander and Jerry Fahrni, PharmD. www.hospitalrx.com
-- Issue 30 --