FDA Approvals Show a Steady Drift Toward More Distinctive Drug Names
Drug names have become stranger on the page. In a retrospective analysis spanning four decades of FDA approvals and related naming records, proprietary names for new molecular entities grew steadily more orthographically distinctive, with longer stretches of uncommon letter combinations and less resemblance to existing names. The pattern was visible not only in branded products but also, to a lesser degree, in trademarks and generic names.
The study examined 1,321 US Food and Drug Administration new molecular entity names from 1985 through 2024, 2,769 FDA Orange Book trade names from 1982 through 2026, 177,337 US pharmaceutical trademarks from 1980 through 2026, and 1,273 generic names from 1982 through 2025. Each name was scored for character trigram surprisal, a measure of how unexpected a three-letter sequence is relative to ordinary English and prior names, and for edit-distance similarity to the closest existing names. Higher surprisal meant a name contained more unusual letter sequences; lower similarity meant it looked less like names already in circulation.
The central result was a consistent increase in name distinctiveness over time. Mean trigram surprisal for new molecular entity names rose in every decade from the 1980s through the 2020s, from 7.21 to 8.50 nats, an 18% increase. The time trend remained strong in linear regression, with b = 0.038, p < .001, and R² = 0.34 after controlling for name length. Nearest-neighbor similarity fell after the early 1990s, reinforcing the same direction of travel from a different statistical angle.
Half of names approved in the 2020s contained a letter sequence found in no common English word. In the 1980s, that figure was 7%.
Namespace Crowding, Branding Strategy, and the Search for Separation
Several forces plausibly sit behind this shift. The naming pool is crowded, and the risk of confusion grows as more products enter the market and more approvals accumulate across therapeutic classes. Regulatory pressure to avoid look-alike, sound-alike names likely pushed manufacturers to move away from conventional spelling patterns and toward forms that are visually and phonetically harder to mistake for existing products. Branding strategy probably amplified that effect, because a proprietary name that is easier to defend legally and harder to confuse with a rival can also be more memorable in a marketplace saturated with similar prefixes and suffixes.
The trademark data tracked the same broad movement. Across 177,337 pharmaceutical trademarks in International Class 005, trigram surprisal also rose over time, with b = 0.016 and p < .001. Generic names changed more slowly, which fits the practical reality that nonproprietary naming is constrained by international naming conventions, pharmacologic stem systems, and the need to preserve recognition across markets. Their trend was still present, but much weaker, with b = 0.011 and R² = 0.063.
Orthographic distinctiveness is not a neutral stylistic flourish. In medication safety, a name that is more unusual may reduce one kind of error while creating another. A highly distinctive string may be less likely to be confused with a nearby product in a database or on a prescription pad, yet unfamiliar letter sequences can make spoken communication harder, slow rapid recognition, and increase the chance of transcription errors, mispronunciation, or incorrect recall. The bedside consequence is straightforward: a name that “stands out” can still fail when it is spoken quickly, scanned inattentively, or entered in a noisy clinical workflow.
Distinctiveness Is Not the Same as Safer Medication Use
The analysis stops short of proving harm, and it does not measure dispensing errors, prescribing mistakes, or patient confusion directly. Its value lies in documenting a long-run linguistic change in pharmaceutical nomenclature and showing that the change is measurable rather than anecdotal. The pattern is large enough to matter for regulators, manufacturers, pharmacists, informaticians, and anyone designing drug databases or clinical decision support systems.
A few implications follow from the data itself. First, naming systems have become more constrained over time, which means future proprietary names will likely continue to compete for a shrinking pool of “safe” and available letter patterns. Second, the rise in unusual orthography may create a false sense of security if distinctiveness is treated as a proxy for safety. Third, medication systems may need to account for the cognitive burden imposed by names that look intentionally irregular, especially in settings where orders are rushed and interruptions are common.
The preprint also has the limitations expected of a retrospective naming study. It is based on publicly available records and computational text metrics rather than direct observation of clinical error. It measures spelling-based distinctiveness, not pronunciation, patient comprehension, or downstream adverse events. The work is also a preprint and has not undergone peer review, so its methods and interpretation should be read with that status in mind.
Still, the underlying signal is hard to ignore. Over four decades, approved drug names moved away from ordinary English and toward more unusual letter patterns, and they did so in parallel across branded names and trademarks. In a system where safe medication use depends on rapid recognition, accurate transcription, and shared understanding across prescribers, pharmacists, and patients, increased distinctiveness may solve one problem while quietly creating another.