Researchers recently presented evidence to suggest that drug name confusion may be more common than previously thought, but that use of alerts in computerized prescriber order entry systems may help prevent them.
By Patricia McTiernan, MS
The Centers for Education and Research on Therapeutics (CERT) program, sponsored by the Agency for Healthcare Research and Quality, aims to increase drug knowledge and awareness, provide clinical information to patients and medical practitioners, and improve the quality of patient care while simultaneously cutting costs. Tools for Optimizing Medication Safety (TOP-MEDS) is the theme of one of only seven CERTs nationwide, and its researchers are working on four main projects that focus on different aspects of medication safety.
In one of the most recent phases of their work, the TOP-MEDs researchers presented evidence to suggest that drug name confusion may be more common than previously thought, but that use of alerts in computerized prescriber order entry systems (CPOE) may help prevent such errors. This article summarizes their most recent findings.
Background on Drug Name Confusion
Bruce L. Lambert, PhD, principal investigator of the TOP-MEDS CERT, says prior research indicates that wrong-drug errors occur at the rate of about 1 per 1,000 prescriptions in both the inpatient and ambulatory settings.
Drug name confusion—a subset of wrong-drug error—is when a clinician confuses the names of two drugs that sound alike or look alike in text.
According to Dr. Lambert, drug name confusion is thought to be the most common type of wrong-drug error, and it can be costly and devastating. For example, Fosamax (a bisphosphonate used to treat osteoporosis) can easily be confused with Flomax (an alpha blocker most commonly used to treat benign prostatic hyperplasia). Hydroxyzine (often used as a sedative) can be confused with hydralazine (a vasodilator used to treat high blood pressure).
As Dr. Lambert points out, there are few effective methods to counteract wrong-drug errors at the point of prescribing. Tall-man lettering (the use of capital letters to distinguish the unique part of the name) has produced mixed results in experiments. Bar coding can be effective at finding wrong-drug errors, but to maximize its benefit it needs to be used at every stage of the drug use process.
The TOP-MEDs team created a clinical decision support (CDS) structure by combining information about drug names, drug indications, and diagnoses in the CPOE system. They tested it in the University of Illinois Hospital and Health Sciences System, in both inpatient and clinic settings. Previous phases had shown that indication alerts can intercept wrong-patient errors (i.e., creating an order in the wrong chart). The more recent phase looked at wrong-drug errors, resulting in intercepting errors at a rate of 1.4 drug name confusion errors per 1,000 alerts.
back to top
Background on the CDS System
Maintaining accurate problem lists in electronic medical records is required by The Joint Commission and is a requirement for meaningful use of health IT as defined by the federal government. Previous research by the TOP-MEDs team used the problem list to identify wrong-patient drug errors.
William Galanter, MD, PhD, MS, medical director of the TOP-MEDs CERT, associate chief health information officer at University of Illinois Hospital and Health System, and a practicing internist, led the development of a list of indications for a subset of drugs in the system’s formulary. The drugs chosen were those that had a relatively small number of indications (for example, metformin, beta blockers, SSRIs). Alerts were set to go off if a physician ordered a drug with an indication that did not match a problem in the patient’s problem list. The physician could then either 1) add the problem to the problem list; 2) change the order; or 3) override the alert.
The team examined more than 125,000 alerts that occurred over a six-year period. According to Dr. Galanter, they determined that an error had been intercepted if the following conditions were met: an alert triggered; the initial medication order was not completed; and the same prescriber ordered a similar-sounding medication for the same patient within five minutes. Similarity of drug names was based on standard measures, and two clinicians performed chart review to determine whether the first, uncompleted order had a plausible indication for use.
The results showed 1.4 drug name confusion errors were intercepted per 1,000 alerts. “From the literature, we had a sense that the rate of wrong drug errors was 1 in 1,000 orders, so I was hoping that if we had a robust intervention, we might see half of that or a quarter of that,” says Dr. Galanter.
As he and Dr. Lambert point out, their team examined only instances where alerts occurred, so they cannot say for certain what the overall error rate is. The alerts were set to go off only for the list of drugs for which indications has been identified. Theoretically, many more drug-name confusion errors may have gone undetected. Still, these latest results support implementing clinical decision support in CPOE to prevent wrong-drug errors, just as previous research showed alerts can also prevent wrong-patient errors and that they can help clinicians populate the problem list.
“What these results show is the potential power of trying to connect indications, diagnoses, and medications in the electronic medical record,” said Dr. Galanter. “Trying to look at those three pieces of information is probably a very fruitful area for future research.”
This research was published in the July 2014 issue of PLOS One. View the study paper online at http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0101977.
Does your organization use the electronic medical record to flag potential drug-name confusion errors? Comment on this post below.
back to top
Patricia McTiernan is editor of the P.S. Blog. Contact her at email@example.com.