On Friday, November 20, as a guest of Humanities Computing colloquia, Sandra Gabriele (professor of Design, York University) presented “Visualization Differentiation in Look-alike Medication Names: Evaluating Design in Context”.
The problem that inspired Gabriele’s study is a troubling one: 7.5 % of patients admitted for acute care experience one or more adverse events; 24% of these are drug-related. Meaning that, all too often, the wrong medication is administered to patients. Why?
Gabriele identified two sources for design errors in hospital drug-selection:
- orthographic similarities of drug names
- phonetic similarities of drug names
Drug names come in two varieties: the generic name (or type), and the brand name (or unique name). Gabriele showed us examples of how medication is stored in hospital pharmacies, presenting pictures of uniform bins of drugs organized alphabetically by name, usually regardless of its intended purpose. One bin contained similarly named blood pressure medications, one for high blood pressure, one for low blood pressure, their names orthographically similar and the labels uniform as well; as a layperson, certainly, I would have been unable to tell the difference at a glance.
Gabriele’s project was to find better ways of designing drug labels for hospital pharmacies. What was most interesting to me was the framework she chose in approaching this problem, asking what was required in effective drug-labelling:
1. Attention: that is, what makes the label distinctive. Some of the designs she used in user tests were changing the colour and weight of the text, or using white text on solid black. User tests showed drug names that were printed as white text on solid black made for the most attention-getting label.
2. Perception: or, legibility issues (cutting down the possibility of confusing orthographically similar drug names), establishing a visible hierarchy of data included on the label, and visual cueing (“chunking”, typographic styles, spatial cues, and mark cues). Gabriele proposed to change the font, so that there was a clearer distinction between upper and lowercase letters, and cleaner font weight. Interestingly enough, users in her test group responded negatively to this change; most drug labels use a Tallman font (which does pose legibility issues like those mentioned), and it seemed that the users (all hospital nurses) were conditioned to using it, when a layperson would have had more difficulty determining minor differences in names. There seems to be some debate over the use of Tallman; a 2006 study in Glasgow showed indicated that Tallman was actually more effective in reducing name-related errors when selecting drugs (Filik et al.).
3. Understanding: Making sure a user can identify and understand all the data available on the label at the glance. Gabriele’s presentation did not delve too deeply into this part of her study, but I would have found this probably the most interesting step in her research. How do users make sense of the labels? Does the reorganization and stratification of data (in the “perception” stage) make a positive difference for comprehension with the trained professional? It seems like, while errors do sometimes occur, changing labels that would avoid errors for a layperson might in fact cause more errors for someone trained to use the current labels in place.
Works cited
Filik, R., Purdy, K., Gale, A., and Gerrett, D. (2006). “Labeling of Medicines and Patient Safety: Evaluating Methods of Reducing Drug Name Confusion.” Human Factors: The Journal of the Human Factors and Ergonomics Society, 48. pp. 39-47.
No comments:
Post a Comment