Testing the Perception of Demonstration Medication Warning Labels
TimeThursday, April 152:00pm - 3:00pm EDT
DescriptionMany hospitals use demonstration medications during in-situ simulation training because they are less expensive than real medications, yet are similar in appearance and form. Demonstration medications typically contain a saline solution, which makes them less expensive, but also a risk to patient safety if they are accidentally administered instead of real medications (U.S. FDA, 2015; Robyn et al., 2015). Unsurprisingly, there have been cases where demonstration medications have been mistakenly administered, or found in place of real medications (Cohen, 2014). Currently, demonstration medications have distinct labeling to help differentiate them from real medications, often in the form of drug-name prefixes (e.g., “DemoDose” or “Practi-”), and misspelled drug names (e.g., “epinephrn”). However, there are no official warning label standards for demonstration medications and different manufacturers use different warning elements on their products.
We conducted a series of experiments to explore which label elements (text prefixes and graphical elements) are the most effective for distinguishing demonstration medications from real medications. We explored this topic using a perceptual classification paradigm in which participants were shown an image of a single medication on a computer screen and then rapidly classified it as either demonstration or real by pressing one of two keys on the keyboard. Each experiment had two different classification tasks. In the real labels task, pictures of on-market medications were presented and in the simulated labels task, newly designed labels with various warning elements were presented. While the real labels task had high external validity but low internal validity, the simulated labels task was the opposite, with high internal validity but low external validity.
In experiment one, 42 participants indicated whether presented images were of real or demonstration medications by pressing buttons labeled as ‘real’ or ‘demo’. In the real labels task, half of the images were of demonstration medications and half were of the corresponding real medications. Of the demonstration medications, half had the text prefix “DemoDose” and the other half had the text prefix “Practi-”. Participants were significantly faster and committed significantly fewer classification errors with the “DemoDose” prefix compared to the “Practi-” prefix. In the simulated labels task, the “DemoDose” and “Practi-” prefixes were used again, along with three possible graphical warning elements: none, yellow triangle with an exclamation point, or a red circle with a human crossed out. Participants were again significantly faster and more accurate when simulated labels contained the “DemoDose” prefix compared to the “Practi-” prefix. Additionally, error rates were significantly lower on simulated labels that contained either of the graphical warning elements, but only when there was also a text prefix on the label. Graphical warning elements on simulated labels without a text prefix were significantly less effective.
Taken together, the results of experiment one indicate that the “DemoDose” text prefix was particularly effective in both real and simulated medication labels. In the simulated labels task, the addition of either the exclamation point or human crossed out graphical warning element made both text prefixes even more effective, but the graphical warning elements were not particularly effective on their own. Note that in this experiment, participants responded using keys labeled as “real” or “demo” and the “DemoDose” prefix was shown to be most effective. We wondered if there was a priming effect from the “demo” label that was driving the superior performance of the “DemoDose” prefix, which we addressed in experiment two.
In experiment two, 33 participants used buttons that were labeled ‘real’ or ‘practice’ and performed the same task on the same stimuli as experiment one. In the real labels task, demonstration medications with the “Practi-” prefix led to significantly fewer classification errors than medications with the “DemoDose” prefix, though there was no significant difference in response times between the two prefixes. In the simulated labels task, the “Practi-” prefix led to significantly faster classifications than the “DemoDose” prefix, though there was no significant difference in accuracy between the two. Once again, participants committed significantly fewer classification errors when demonstration medications contained graphical warning label elements, but only when one of the two text prefixes was present.
Experiment two indicated that participants may, indeed, have been primed by the keyboard labels in experiment one. When the response choices were labeled as “demo” or “real”, the “DemoDose” prefix had better performance in experiment one. However, when the response choices were labeled as “practice” or “real”, the “Practi-” prefix yielded better performance in experiment two. We are currently collecting data on experiment three, with response choices labeled as “simulated” or “real”. This experiment should reveal classification performance when participants are not influenced by one of the response labels being similar to one of the text prefixes being tested.
Importantly, in both experiments, graphical warning labels were most effective when accompanied by a text prefix in the simulated medication name. When demonstration medications contained both text prefixes and graphical warning elements, error rates were cut by one third to one fifth across the two experiments. There were no significant differences between the two warning label types in either study, indicating that both were effective at reducing error rates.
Demonstration medications are an important element in simulation training, but they have the potential to cause harm if they are accidentally administered to patients. Consistent and effective warning labels for demonstration medications should help reduce medication dispensation errors with these products, but there are currently no standards in the marketplace for such warning elements. The current research indicates that text prefixes improve the recognition of demonstration medications, especially when they are combined with graphical warning elements.