Prospektive Risikoanalyse: Die Ähnlichkeit von Medikamentennamen in der Drugbank-Datenbank

Abstract

In various studies, a frequency of adverse drug events in medical treatments was found ranging between 0.7% and 6.5%. About 12% of these medication errors are due to confusion resulting from similar sounding names (sound-alike) or similar spellings (look-alike). The problem of sound-alike and look-alike drug names, which however should always be a unique name, is monitored by various institutions at international, European and national levels. In order to prevent the confusion of drugs, it is recommended to maintain lists of look-alike and sound-alike drug names (LASA list). These LASA lists, however, reflect only on pairs of drug names which already had been confusedIn this study, we prospectively describe the similarity of drug names based on different algorithms. For this purpose, methods analysing both orthographic and phonetic similarity were applied. The study included 62.354 drug and product names of the open accessible database „Drugbank“. According to our results about 50% of the drug names listed in the Drugbank are characterized by comparable similarity measures as the already confused drug names listed in the LASA-list of the Institute for Safe Medication Practices (ISMP) and the Federal Association of German Hospital Pharmacists e.V. (ADKA). 4.822 name pairs generated from the Drugbank were different only in one letter, while 22.039 name pairs differed in two letters or editing steps (Levenshtein distance). Analysing the phonetic similarity resulted in a resemblance between 20% and 96% of names.

Keywords

• patient safety
• sound alike
• look-alike
• prospective analysis