PHILADELPHIA In the first study of its kind, researchers led by The University of Pennsylvania School of Medicine's Ross Koppel, Ph.D. studied how hospital nurses actually use bar-coded technology that matches the right patient with the right dose of the right medication. The surprising result is that the design and implementation of the technology, which is often relied upon as a "cure-all" for medication administration errors, is flawed, and can increase the probabilities of certain errors.
Equally surprising is that the urgencies of care and the ingenuity of nurses to cope with these shortcomings have the unintended consequences of creating other medication errors. These findings appear in the July/August issue of the Journal of the American Medical Informatics Association (JAMIA). The study also illustrates how adjustments to workflow and the technology can dramatically reduce the risk of these errors.
These barcode systems usually consist of handheld devices and computers that match machine-readable barcodes on patients and medications. If they match, and if they are consistent with the ordered medications, the medications are given. If not, usually a signal goes off telling the nurse of a discrepancy.
The study was conducted at 5 hospitals in the Midwest and on the East Coast, but not at the Hospital of the University of Pennsylvania (HUP) because it does not yet have medication barcoding. Penn's Ross Koppel, Ph.D. and his colleagues from other healthcare systems examined close to a half-million instances where nurses and other staff scanned patients and medications. The researchers found a remarkably high proportion of scans involved nurses overriding the technology with workarounds to compensate for difficulties with the barcode systems. These researchers found that nurses scanning the barcode on the medication or the patient's ID bracelet overrode the technology for 4.2% of patients charted and for 10.3% of medications charted. In contrast, vendors of barcode medication administration (BCMA) systems report error rates that are a small fraction of this study's numbers; but vendors focus primarily on the ability to physically affix and read barcodes, not on the totality of the many processes in actual use. In addition to examining the million scans, Dr. Koppel and colleagues spent years shadowing nurses using the technology, participated in many BCMA implementation meetings, and conducted scores of interviews with pharmacists, nurses, and IT leaders.
Hospital patients, on average, are subject to one medication administration error a day, according to the Institute of Medicine, and in hospitals, medication administration accounts for 26% to 32% of adult patient medication errors. Thus, an automated system using barcodes to reconcile a patient's medications and orders with the patient's identity would be a great advance, helping to ensure the right patient receives the right dose at the right time.
But what Penn's Professor Koppel and his colleagues found in the five study hospitals were 31 "causes" of problems that engendered workarounds by the nurses. These causes included: unreadable medication-barcodes (crinkled, smudged, torn, missing, covered by another label); malfunctioning scanners; unreadable or missing patient-ID-wristbands (chewed, soaked, missing); non-barcoded-medications; medications in distant refrigerators, lost wireless connectivity; problems with patients in contact isolation, and emergencies. In some cases, if the pharmacy sent two 10mg tablets for a 20mg order, the scanners/computers would not accept the medications. Nurses devised workarounds to compensate for the awkward and inconvenient aspects of the barcode technology. These nonstandard procedures consisted of, for example, affixing extra copies of patient ID barcodes on desks, scanning machines, clipboards, supply room, and doorjambs, as well as carrying several pre-scanned patient's medications on one tray. Ross Koppel, Ph.D., Center for Clinical Epidemiology and Biostatistics at the University of Pennsylvania School of Medicine and the Sociology Department at the University of Pennsylvania, emphasized that "It's not that staff are lazy or careless, it's that the system does not work as well as it should. If the refrigerated medication is two floors and a long hallway away, you're not going to wheel your 87 year old patient to the fridge. You make a copy of her barcode. And while you do that, you help another two patients who also need refrigerated medications."
"Bar-coding is still under development," says Koppel. "Administrators and vendors may expect it to be fool-proof, but users know it's not. It's a very promising technology that still requires constant refining and careful observation of on-the-floor workflow to get it right."
The researchers found that in the pressurized, "can-do" culture of today's hospital, nurses compensated for the imperfect technology and workflow by devising 15 types of workarounds. The study also presents typologies of workarounds, BCMA "causes," and the kinds of errors associated with each.
Every day BCMAs save lives and stop errors, says Dr. Koppel, and the published study documents thousands of medication errors avoided via these systems. In addition, the article also lists many recommendations for identifying the problems and mitigating workarounds. Four of the study hospitals reduced the number of overrides dramatically by following these recommendations.
"The causes of workarounds are neither rare nor secret," added Koppel. "They are hidden in plain sight, obscured by, among other things, a blind faith in technology and the urgent needs of patient care. Clever as they are, workarounds are the unintended consequence of a technology in need of continuing and in situ evaluation."
Professor Ross Koppel's research on healthcare information technology (HIT) came to national prominence a few years ago with a JAMA article on medication errors associated with computerized physician order entry systems (CPOE). But he has published widely on HIT, noting its many benefits as well as its problems. In the same July/August issue of JAMIA, in fact, he is also the first author on a paper that presents a way of using CPOE to detect and prevent medication errorsa decidedly pro-CPOE piece. Koppel remarked, "Many vendors and their supporters mistakenly believe I'm some sort of Luddite. That's the exact opposite of my position. I view these technologies as vital; that's why we must make them work to help clinicians and patients. Right now, the vendors and true believers focus on marketing HIT, and attack any criticisms as anti-technology. That's the worst way to improve these essential tools."
This study challenges assumptions of how these increasingly popular bar-coded medication administration systems are actually used in hospital practice. "It is not enough to tell the staff to "do it right", concludes Koppel, "rather repeated examinations and corrections of the technologies in actual use will help optimize their roles in preventing medication error and enhancing patient safety."
As noted, HUP does not use barcode medication administration systems, but requires staff to manually check identifiers to correctly match patient and medication. When HUP implements a medication barcode system, it will incorporate the findings from this research.
|Contact: Marc Kaplan|
University of Pennsylvania School of Medicine