Effects of Text Enhancement on Reduction of Look-Alike Drug Name Confusion: A Systematic Review and Meta-analysis

Abstract

Supplemental Digital Content is Available in the Text. Background and Objectives: Drug name confusion induced by look-alike drug names represents a serious health care management problem in practice. Text enhancement by changing visual attributes of look-alike drug names has been proposed and widely applied in practice to mitigate drug name confusion. However, the effectiveness of text enhancement on reducing drug name confusion is yet to be determined. This study aimed to explore the effects of text enhancement on reduction of confusion caused by look-alike drug names through systematic review and meta-analysis. Methods: We searched 5 databases (from database inception to January 2020) for empirical studies that examined the effects of text enhancement on reduction of look-alike drug name–induced drug name confusion. The pooled outcome data were analyzed using either meta-analysis or a narrative synthesis approach. Results: From the 351 identified articles, 11 articles representing 20 individual trials were included. Five basic text enhancement methods were revealed, including Tall Man, red, boldface, contrast, and size enhancement, from which 4 Tall Man variants and 6 text enhancement combinations were derived. The meta-analysis results showed significant reduction in omission errors when using Tall Man (standardized mean difference [SMD] = −0.628, 95% confidence interval [CI]: −1.018 to −0.238, P = .002), red (SMD = −0.516, 95% CI: −1.002 to −0.030, P = .038), boldface (SMD = −1.027, 95% CI: −1.240 to −0.814, P < .001), and contrast (SMD = −0.437, 95% CI: −0.869 to −0.004, P = 0.048), as compared with lowercase. This finding was also supported in our subgroup analysis by task type for name differentiation tasks. No other significant effects of text enhancement were found for either commission errors or response time. Conclusions: Using Tall Man, red, boldface, or contrast could help reduce omission errors (ie, wrong medication selection) caused by look-alike drug names, particularly in name differentiation tasks. However, no text enhancement could shorten name search and/or differentiation time. Our findings could facilitate the understanding of the effects of text enhancement in the prevention of confusion errors caused by look-alike drug names and promote the application of text enhancement in practice.