Most medication mentions (64.0%) in the progress notes analyzed were generic names. Generic mentions are more likely if the note author is a pharmacist or medical student, if the medication has a generic version available, if the note is targeted toward healthcare workers rather than patients, if the medication has been on the market for longer than 20 years, and if the generic name is not significantly harder to say or write than the brand name. Each of these factors plays a role, but most of these effects are relatively small. Individual clinicians’ documentation habits appear to play a much larger role, with attending physicians using brand names anywhere from 3.7%-73.3% of the time. It is likely that more granular properties such as surrounding words and sentence context also affect brand name usage. In short, brand name usage is a complicated phenomenon influenced by a wide variety of note-level, writer-level, and medication-level factors. Our study represents the largest study of its kind to date, both in terms of the number of medications analyzed and the number of notes analyzed.
Even after removing templated medication lists from our analysis, a majority of mentions were generic. This is primarily a function of the documentation habits of physicians, who were the most common source of medication mentions (61.9%) in our data set. The proportion of brand mentions in our data set is smaller than in past studies[10, 12, 13] notably, Steinman’s study of 2003 U.S. outpatient practice had a median brand name mention frequency of 98%. This may be a result of our analysis including many common medications that have been generically available for decades (e.g., aspirin, heparin, antibiotics, etc.) - many clinicians likely do not even know the old brand names. Furthermore, the share of generic prescriptions in the U.S. has increased dramatically over the past few decades,[15] reflecting increased cost-consciousness as well as older brand-name medications going off patent. Ouyang et al.’s study analyzed text pages, a more informal mode of communication than medical documentation, and one that is often used for urgent or emergent communication. The informality or urgency of the communication medium may increase the likelihood of brand name usage. Lastly, awareness of pharmaceutical industry influence may have improved over the past decades. Within our system, brand name mentions have been decreasing steadily year-over-year from 2015 to 2020 (45.7–31.9%). This effect, over such a short time-scale, is unlikely to be explained by a significant change in the proportion of actual generic vs. brand-name prescriptions, and more likely reflects alteration in documentation habits. Still, there is a sizable proportion of brand name usage even within the formal context of medical documentation (~ 36%).
One notable finding is the large difference among pharmacists, physicians, and nurses in brand name usage. This corroborates previous work demonstrating that pharmacists are significantly less likely (and nurses more likely) on average to use brand names than physicians.[12] This difference may be a function of different educational focuses among healthcare workers. Another possible explanation is the degree of patient contact and education each healthcare professional provides. Numerous studies have shown patients tend to have a preference for branded versions of drugs over generics,[5–6] and may be more familiar with brand names due to their easier pronunciation and the influence of direct-to-patient advertising. This point is supported by our findings related to patient instructions, which are the only note types targeted toward patients and have a significantly higher brand name mention frequency. Since nurses typically spend more time with patients than physicians and pharmacists, they may be more accustomed to use brand names in patient education. Residents are more likely than medical students to use brand names, and attendings more likely than residents, suggesting that as clinicians progress through their training, their brand name usage increases. This may be due either to their increased contact with patients or the modeling of behavior after senior colleagues.
Our study demonstrates that brand name usage is prevalent in clinical documentation. There are legitimate concerns about brand name usage, and multiple possible remedies to reduce this problem. On the individual level, the wide variation in brand name usage suggests a role for targeted educational intervention. We can provide education about the reasons to avoid brand names and build electronic tools (reminders or automatic corrections) to facilitate generic usage. Given the evidence for the potential influence of patient preference, re-examining the benefits and harms of direct-to-patient advertising (DTPA) may also be useful.
Our study has limitations. Notably, we did not study combination drugs with multiple active ingredients and, given the scope of the study, omitted medication abbreviations. Some medication mentions, such as those created using templates or copied forward from prior notes, may not reflect the writer’s individual choices. Our study is limited to official medical documentation, which may not reflect broader habits in verbal communication or clinician cognition. Our study involves a single academic healthcare system in the United States, and may not generalize to other healthcare settings.