Researchers: Yuval Bitan¹

Background: Errors during medical treatment occur at a high rate and are considered to be one of the leading causes of death in developed countries. One of the most common types of medical errors is medication error.

Objectives: This study investigates the use of visual patterns as a means of improving medication identification when providing medicine. We developed labels divided according to the spatial frequency index of the background, and we used them for an experiment to test if the patterns were able to provide performance improvements beyond text labels in a search and recognition task.

Method: The experiment compared three types of labels: white labels and two types of patterned labels. We designed 75 labels of two types, 30 white and 45 patterned labels, which included 30 low spatial frequency labels and 15 high spatial frequency labels. The experiment included 60 participants where 20 participants experienced each of the label types.

Findings: Our results reveal significant differences between use of the labels both in response time and accuracy. The average response time for the patterned labels was 2,810 ms with the low spatial frequency labels, and 3,477.62 ms with the high-low spatial frequency labels, compare to 4,410.56 ms with the white labels (p<0.05). The accuracy of the patterned labels was 80.5% and 74.7% with the white labels. We also found differences in the search pattern for the patterned labels compare to the white labels.

Conclusions: Results indicate that adding patterns to the background of text labels significantly reduced search times, and improved accuracy. These results suggest that patterned labels can reduce search time and may reduce the likelihood of errors.

Recommendations: The use of visual patterns as background that were evaluated in this study present promising potential for a new approach in improving medication identification. Our recommendations suggest that policy makers should consider the use of such solution as part of future standard for medication labeling.

Research end date: 08/2020