In our discussion of the article by Kesselheim and coauthors, the editors debated the clinical implications of the findings. Some noted that the association between color change and nonpersistence with generic medicines was small. Color discordance preceded 1.20% of episodes of nonpersistence compared with 0.97% of episodes of persistence. This seems like a small effect: an absolute difference of 0.23%, slightly less than 1 out of 400.
However, others noted that over time, this seemingly small effect size could become quite important. Most patients refill a medicine at least 4 times a year. Over the course of several years, the risk of nonadherence owing to color change can add up even for 1 medicine. And consider the very common situation of an older patient dealing with a confusing regimen of many medicines. If they are taking 9 medicines, they get at least 36 “opportunities” a year to experience a color change. This seemingly small risk then starts to appear very substantial. Throw in some cognitive impairment and other sources of confusion, such as changing medication regimens each time a patient transitions in and out of the hospital, and it becomes easy to imagine how a medication color change can lead to a significant medication-related adverse event.
Covinsky KE. Debating Effect SizesComment on “Variations in Pill Appearance of Antiepileptic Drugs and the Risk of Nonadherence”. JAMA Intern Med. 2013;173(3):209. doi:10.1001/jamainternmed.2013.1545