Researchers Get $1.6 Million NIH Grant to Help End Medication Errors Caused by Similar-Sounding Drug Names

BUFFALO, N.Y. -- The National Institute of Medicine estimates that medication errors result in more than 7,000 fatalities a year and a 2002 FDA report claimed that in the United States alone, medication errors made throughout the drug distribution system injure another 1.3 million patients a year.

Confusing nomenclature, labeling, packaging and handwriting bear some responsibility, but thousands of drugs with confusing look-alike and sound-alike names account for one-fourth of all reported errors.

A look at some of the most commonly prescribed drugs in the pharmaceutical arsenal suggests the extent of the problem: Vioxx/Videx, Clinoril/Clozaril, Serzone/Seroquel, Zantac/Zyrtec, Toradol/Torecane, Zoloft/Zocor, Clonidine/Klonopin, Celebrex/Celexa.

Paul Luce, professor and chair in the Department of Psychology in the University at Buffalo's College of Arts and Sciences, is co-principal investigator on a new study that will look at auditory perception of drug names and develop a software program that will address the problem.

Bruce Lambert, associate professor of pharmacy administration at the University of Illinois, Chicago, is principal investigator on the five-year study, funded by a $1,678,049 grant from the Agency for Healthcare Research and Quality (AHRQ) of the National Institutes of Health.

Luce, who directs UB's Language Perception Laboratory, says the long-term objective of the study is to minimize the overall incidence of name-confusion errors related to drugs. Its short-term goal is to develop an empirically validated, user-friendly software tool that can be used to screen proposed drug names against databases of existing drug names.

For several years, pharmacists, linguists and other scientists have studied the danger posed by similarities between the names of existing drugs and those assigned to new drugs.

Previous analysis by Lambert and other researchers, for instance, described a variety of distributional characteristics of drug names that cause confusion, including the number of characters per name, the number of syllables per name and the number of words per name. The new study will employ perceptual models developed by Luce to look at problems with the auditory recognition of similar-sounding medications.

Although drug companies and regulators screen names prior to approval, Luce says the screening process is error-prone because of an over-reliance on subjective assessments of auditory similarity between names. In this study, he says, the relative confusability of a word will be based on validated, objective criteria derived from studies of clinicians' and lay persons' actual auditory perceptual errors.

Although it might be supposed that lay people have more difficulty differentiating among the drug names, Luce said the researchers expect to find problems with auditory perception and identification to be the same among adult lay people, physicians, nurses and pharmacists.

Psychologists know that the ability to recognize a particular word when heard is affected by the intelligibility of that word. Luce's Neighborhood Activation Model (NAM), which will be used as the theoretical framework of the study, holds that auditory perception errors depend not only on the intelligibility of the target word, but on the similarity and frequency of words in its perceptual "neighborhood."

This prediction is embodied in Luce's Frequency-Weighted Neighborhood

Probability Rule (FWNPR), which holds that words with many phonetically similar neighbors will be identified more slowly than words without many neighbors.

Using the outcomes of auditory perceptual identification tests, the researchers will develop a user-friendly software tool that will predict accurately the probability that a specific name will produce auditory confusion.