The Quiet Agreement Not to Notice: Why Tardive Dyskinesia Still Goes Unnamed

Mara is a 58-year-old woman with a 20-year history of schizoaffective disorder, stable for the past six years on risperidone 4 mg daily. She comes in for a routine medication check, and everything looks fine on paper: mood stable, no hallucinations, adherent to her regimen, working part-time at a bakery. Her clinician almost wraps up the visit early. Then, while Mara is talking about a coworker, her clinician notices it — a faint, irregular pursing of her lips, a tongue that seems to be pushing gently against her cheek every few seconds. Mara hasn't mentioned it. When asked, she says, “Oh, that? I don't even notice it happens.” 

Mara is a composite, not a real patient — but the moment is a familiar one, and it's worth sitting with instead of rushing past. Because what happened in that room wasn't really a knowledge gap. Her clinician knows what tardive dyskinesia (TD) looks like. What almost didn't happen was someone deciding to say something out loud. That's the part of TD that clinical training under-covers: not the phenomenology, but the quiet incentives both people in the room have to leave it unnamed. 

Two people, two reasons to stay quiet 

Patients frequently don't report TD movements — not always because they can't feel them, but because orofacial and limb movements often escape conscious monitoring, especially once they've been present long enough to become background. Add in embarrassment: TD is visibly stigmatizing in exactly the way patients are often working hardest to overcome, whether that's holding down a job, dating, or simply not standing out in public. Bringing it up can feel like drawing attention to something they'd rather minimize, even to themselves. 

Clinicians have their own reasons to look past it. The condition is, as psychiatric literature has bluntly put it, one that prescribers are “understandably loath to diagnose”1 — because it's iatrogenic, because there's a reflexive worry about liability (“did I document informed consent clearly enough when I started this medication years ago?”), and because for decades there wasn't much to offer once it was named. That last part is now outdated — FDA-approved treatments have existed since 2017 — but the older instinct to under-name TD because naming it felt like opening a problem with no solution hasn't fully caught up to that fact. A 2025 review of electronic health record data found TD movements are documented in clinical notes considerably more often than TD is actually coded as a diagnosis — evidence that clinicians are seeing it and choosing, consciously or not, not to formalize it.2 

Mara's case is really a case of two people arriving at the same unspoken agreement: let's not make this a thing. 

The conversation happens twice — get the first one right 

That agreement doesn't start in the exam room years into treatment. It often starts earlier, in a different conversation clinicians describe struggling with just as much: how to disclose TD risk when starting an antipsychotic without either glossing over it or handing an already frightened, possibly unstable patient a reason to refuse treatment they need. 

It's a real tension, not an imagined one — but it's worth knowing that the evidence doesn't support the fear driving it. In one of the few studies to test this directly, patients who received a formal, standardized informing process about TD risk retained significantly more knowledge about it at six-month follow-up than patients who weren't formally informed — with no increase in nonadherence, psychiatric admissions, or need for higher doses in the informed group.3 Separately, when patients with lived experience of TD have been asked about their own informed consent process, the complaint runs the other direction: many report feeling they weren't told enough to make a real decision, not that they were told too much.4 The instinct to soften the disclosure to protect adherence is understandable, but it isn't backed by what happens when clinicians test it. 

A few things seem to make the conversation land better in practice: anchoring it in concrete terms rather than vague ones (“a movement disorder that becomes more likely the longer you're on this medication” lands differently than reciting “tardive dyskinesia” as one line on a list of side effects); explicitly naming that effective treatments now exist if it develops, which removes the fatalism that made this conversation feel like delivering a life sentence a generation ago; treating it as an ongoing conversation revisited over the first months of treatment rather than a single disclosure buried in an initial consent form; and, where appropriate, including a family member or support person, which both reinforces the information and gives the patient another set of eyes down the line. 

That last point connects back to everything above. A patient who was told early, clearly, and without alarm that this is a known, monitored, treatable risk is a patient who's more likely to mention it later instead of minimizing it the way Mara did. The first conversation is what makes the second one possible. 

Why “let's just keep an eye on it” isn't a neutral choice 

The instinct to defer — reduce the dose a little, switch agents, monitor at the next visit — feels conservative and low-risk. It usually isn't. TD is frequently persistent even after the causative medication is stopped or changed: in one retrospective study of 108 patients seen at a university movement disorder clinic after their dopamine-blocking drug was permanently withdrawn, only 13% achieved complete resolution of their tardive syndrome — fewer than 3% from withdrawal alone — and no clinical factor, including age, baseline severity, or duration of exposure, predicted who would be in that minority.5 Watchful waiting isn't a pause button. For most patients, the movements that are present today are likely to still be present next year, regardless of what happens with the antipsychotic. 

That reframes the clinical stakes considerably. The choice isn't between “act now” and “act later with no cost.” It's between treating a window that's open now and accepting that the same conversation, if delayed, will likely need to happen anyway — just with a patient who has lived longer with symptoms that didn't need to go untreated that long. 

The gap that matters most isn't recognition — it's what happens after 

Here's the number that should reframe this conversation: in a large claims-data analysis of adults on antipsychotics between 2017 and 2022, diagnosed TD prevalence rose modestly from 0.45% to 0.57% — but VMAT2 inhibitor use, while it also rose, sat at just 0.22% by the end of that period.6 Diagnosis is necessary, but it isn't the finish line, and the data suggests a meaningful share of patients who get the diagnosis still don't get started on a first-line treatment. That's a different problem than “clinicians miss TD,” and it points to a different fix: a diagnosis that doesn't trigger a treatment conversation is functionally the same as no diagnosis at all, from the patient's point of view. 

This matters because the treatments work. Real-world data collected from clinicians prescribing VMAT2 inhibitors found symptom improvement in roughly 90% of patients, with the majority also reporting meaningful gains in social and occupational functioning — the things patients actually came in worried about, even if they never said “tardive dyskinesia” out loud.7,8 

What actually breaks the silence 

None of this requires new instruments so much as a shift in what the instruments are for. The AIMS exam, done properly — observing at rest, during physical and cognitive activation, and while the patient is distracted in conversation rather than performing for the exam9 — exists to catch what neither party would otherwise say. The DSM-5 timing criteria (onset after at least three months of exposure, or one month in patients over 60, with movements persisting beyond four to eight weeks)10 exist to give clinicians a defensible, non-arbitrary reason to name what they're seeing rather than waiting for more certainty that may never come. And knowing the movement vocabulary — TD's irregular, non-rhythmic, choreiform quality versus the rhythmic resting tremor and rigidity of drug-induced parkinsonism — matters not just for accuracy, but because misreading one as the other can lead to an anticholinergic that makes TD worse, reinforcing the impression that nothing helpful can be done.10,11 

Used this way, these tools aren't a checklist for catching TD. They're a way of giving clinicians permission to say the thing they already suspect, and a way of giving patients a reason to stop minimizing what they've already noticed. 

The takeaway 

The knowledge needed to identify TD has been widely available for years, and most clinicians reading this already have it. What moves the needle now is smaller and harder: treating “I don't want to make this a bigger deal than it is” — in ourselves or in our patients — as the actual barrier, and treating a diagnosis as the start of a treatment conversation rather than the end of the visit. Mara's tongue movements didn't change her diagnosis or her adherence. But her clinician naming them, instead of letting the moment pass the way it almost did, is the only part of that visit that could have changed what happens to her next. 

References 

1. Goldberg JF. Tardive dyskinesia: assessing and treating a debilitating side effect of prolonged antipsychotic exposure. Psychiatric News. 2021. https://psychiatryonline.org/doi/10.1176/appi.pn.2021.3.10 

2. Identifying the diagnostic gap of tardive dyskinesia: an analysis of semi-structured electronic health record data. BMC Psychiatry. 2025;25(1):407. 

3. Munetz MR, Roth LH. Informing patients about tardive dyskinesia. Am J Psychiatry. 1985;142(7):866-871. 

4. Bergman H, Walker DM, Nikolakopoulou A, et al. Systematic review of interventions for treating or preventing antipsychotic-induced tardive dyskinesia — Patient and Public Involvement report. Health Technol Assess. 2017;21(43). https://www.ncbi.nlm.nih.gov/books/NBK447493/ 

5. Zutshi D, Cloud LJ, Factor SA. Tardive syndromes are rarely reversible after discontinuing dopamine receptor blocking agents: experience from a university-based movement disorder clinic. Tremor Other Hyperkinet Mov (N Y). 2014;4:266. doi:10.5334/tohm.199 

6. The use of VMAT2 inhibitors for tardive dyskinesia. J Clin Psychopharmacol. https://www.ovid.com/jnls/psychopharmacology/abstract/10.1097/jcp.0000000000002139 

7. Chart extraction/clinician survey shows symptom impact and favorable treatment outcomes with VMAT2 inhibitors in patients with tardive dyskinesia. Neurology. 2020;94(15 suppl). https://www.neurology.org/doi/10.1212/WNL.94.15_supplement.14 

8. Real-world use and impact of VMAT2 inhibitors in patients with tardive dyskinesia. MDS Abstracts. 2020. https://www.mdsabstracts.org/abstract/real-world-use-and-impact-of-vmat2-inhibitors-in-patients-with-tardive-dyskinesia/ 

9. Guy W. ECDEU Assessment Manual for Psychopharmacology, Revised. DHEW Publication No. ADM 76-338. Rockville, MD: National Institute of Mental Health; 1976. 

10. Ward KM, Citrome L. Antipsychotic-related movement disorders: drug-induced parkinsonism vs. tardive dyskinesia—key differences in pathophysiology and clinical management. Neurol Ther. 2018;7(2):233-248. https://link.springer.com/article/10.1007/s40120-018-0105-0 

11. Differentiating tardive dyskinesia from similar drug-induced movement disorders. NeurologyLive. https://www.neurologylive.com/view/differentiating-tardive-dyskinesia-similar-drug-induced-movement-disorders