Blog — Clinical Spanish

Spanish for neurology nurses: the last-known-well question that changes the tPA calculation, the NIHSS arm drift that disappears when you test too quickly, and the seizure witness history that arrives as “se cayó y no me respondía”

Daniel Morales was 58 years old. His daughter Rosa found him at the kitchen table at 7:30 in the morning, unable to say his own name clearly. He told her “me siento raro” when she asked what was wrong. He could not explain more. Rosa called 911. At the emergency department, the triage nurse asked when he had last been well. Rosa said “estaba bien anoche.” The nurse documented “last known well: last night.” It took twelve minutes and the stroke coordinator’s specific questioning to establish that Daniel had been asleep since 10:30 PM, that Rosa had not seen him awake and neurologically normal this morning before calling, and that “anoche” therefore meant 10:30 PM—not “this morning before I called,” which was what the triage documentation had implied. The difference was eight hours. By the time the window was confirmed, it was already gone.

Rosa’s “anoche” and the eight-hour window

Daniel was right-handed. He had been a landscaper for thirty years. His wife had died four years earlier. He lived alone now, but Rosa, his daughter, lived three blocks away and checked in most mornings. He was 58, with hypertension and type 2 diabetes, last A1c 8.4 in February.

Rosa called 911 at 7:42 AM. She met the paramedics outside and rode to the ED with him. At triage, the nurse asked: “¿Cuándo empezaron los síntomas?” Rosa said she had found him like this when she arrived at 7:30. The nurse asked: “¿Cuándo estuvo bien por última vez?” Rosa said: “Estaba bien anoche.”

The nurse documented “last known well: last night (per daughter).” That documentation, without a clock time, is clinically useless. “Last night” could mean 9 PM after dinner. It could mean midnight when he texted Rosa goodnight. It could mean 2 AM if Daniel had insomnia and had been awake and functional in the early hours.

What “anoche” actually meant, in this case, was that Rosa had spoken to her father on the phone at 10:30 PM—he had sounded fine, said he was going to bed, and that was the last she knew of him until 7:30 AM. He had been asleep from 10:30 PM to some unknown time in the night, and either woke symptomatic or became symptomatic during sleep. Rosa had not seen him before calling 911. She had heard him make a sound in the kitchen and found him at the table.

The tPA window from a wake-up stroke, in current clinical practice, uses the last known well as the anchor—not the time of discovery. If last known well was 10:30 PM and the patient was found at 7:30 AM, the elapsed time at presentation is nine hours. The extended window protocols require imaging selection criteria (DWI-FLAIR mismatch) that Daniel’s facility could obtain, but the twelve minutes spent clarifying what “anoche” meant were twelve minutes that reduced the margin for everything that followed.

The stroke coordinator who established the true timeline did so in four questions that the triage nurse had not asked.

Three failure modes for neurology nursing in Spanish

1. The last-known-well question where “anoche” is not a time

The last-known-well (LKW) is the most clinically consequential piece of information in an acute stroke evaluation. Everything downstream—tPA eligibility, mechanical thrombectomy window, imaging protocol selection, transfer decision—depends on having a specific clock time, not a relative term. The failure mode is accepting a relative term and documenting it as if it were a time.

“Anoche” is not a time. “Esta mañana” is not a time. “Antes de dormirse” is not a time. “Cuando lo vi normal” is not a time. Any of these answers to “¿cuándo estuvo bien?” requires a follow-up question that produces a clock time.

The question sequence that pins the window:

“¿Cuándo fue la última vez que usted lo vio bien—hablando normal, moviéndose normal, sin nada raro?”

(When was the last time you saw him well—speaking normally, moving normally, nothing strange?)

If the family member gives a relative term, follow immediately with:

“¿A qué hora fue eso, aproximadamente?”

(About what time was that?)

If the patient was asleep and the witness did not see them awake and normal before calling:

“¿A qué hora lo vio bien por última vez despierto—antes de que se durmiera?”

(What time did you last see him well and awake—before he went to sleep?)

The critical distinction that Rosa’s case illustrates: a patient who was asleep from 10:30 PM and was found symptomatic at 7:30 AM has a last-known-well of 10:30 PM—not “this morning.” The stroke may have begun at any point during those nine hours. The witness who says “estaba bien anoche” and then, when asked, confirms the last contact was a 10:30 PM phone call, has just given you 10:30 PM as the LKW. Document the clock time.

Two additional questions that surface wake-up onset when the witness is not sure:

“¿Él se levantó de la cama solo esta mañana? ¿Lo vio levantarse?”

(Did he get up from bed on his own this morning? Did you see him get up?)

“¿Le habló esta mañana antes de llamar al 911? ¿Respondió normal?”

(Did you speak to him this morning before calling 911? Did he respond normally?)

A “no” to either of these, combined with a clear phone call at 10:30 PM, closes the gap: last known well is 10:30 PM.

One phrase that makes things worse: “¿Cree que estuvo bien esta mañana?” (Do you think he was well this morning?) A leading question in either direction—toward “this morning” or “last night”— will be answered in the direction the witness thinks the nurse wants. Ask what they observed, not what they believe.

2. The NIHSS arm-drift item where the test ends before drift begins

The National Institutes of Health Stroke Scale (NIHSS) is administered in English by default in most US stroke workflows, but any nurse doing bedside NIHSS on a Spanish-speaking patient must instruct in Spanish or use a certified interpreter. The arm-drift item (items 5a and 5b) is where language failure most often produces a false-normal score.

Mild hemiparesis in the acute stroke patient does not look like paralysis. It looks like an arm that goes up correctly and then begins to drift downward—not immediately, but over five to eight seconds. A nurse who delivers the instruction, sees both arms go up, and scores 0 after three seconds has not administered the item. She has administered the first three seconds of a ten-second test.

The instruction in Spanish:

“Voy a pedirle que levante los brazos. Levante los dos, así — [demonstrate] — palmas hacia arriba, como si estuviera pidiendo lluvia. Manténgalos así hasta que yo le diga que puede bajarlos.”

(I am going to ask you to raise your arms. Raise both of them, like this — palms facing up, as if you were asking for rain. Keep them like that until I tell you you can lower them.)

Then wait the full ten seconds. Watch both arms simultaneously. Watch for:

• Drift: one arm descends while the other holds. Drift beginning at second 7 is still drift. Score 1 on the drifting arm.
• Asymmetry at the start: if one arm does not come up at all, or only comes up partially against gravity, score 2 or 3 before the ten seconds—do not wait for drift that cannot occur.
• Corrective movement: a patient who starts to drift and then re-lifts the drifting arm has just shown you a drift. Score it.

What not to do during the ten seconds:

“Levante más ese brazo.” — coaching is not permitted after the initial instruction and invalidates the item. “¿Está bien?” — a social question during the hold disrupts attention. Stay silent and watch.

The bilateral comparison is the clinical heart of the item. Asymmetry— one arm holding while the other drifts—is the finding. Bilateral drift (both arms descending equally) is more likely fatigue or reduced comprehension than focal hemiparesis.

Two additional NIHSS items where Spanish instruction matters:

Facial palsy (item 4): the standard instruction is “show me your teeth” or “smile.” The problem: a polite patient will give a minimal social smile that does not stress the facial musculature enough to reveal subtle asymmetry. Better:

“Muéstreme todos sus dientes tan grandes como pueda— como si fuera a aparecer en una foto.”

(Show me all your teeth as big as you can—like you are going to be in a photograph.)

The “for a photograph” anchor consistently produces a fuller facial movement than “sonría,” because it signals that effort is expected.

Speech (item 9): the NIHSS uses a standard passage (“You know how”) and object-naming cards. For a Spanish-speaking patient, use the validated Spanish version of the passage. Do not ask “¿cómo está?” as a speech sample—a one-word social response does not assess paraphasia, anomia, or fluency. Ask for a sentence:

“¿Me puede decir qué hace usted en un día normal—cómo es su día?”

(Can you tell me what you do on a normal day—what is your day like?)

A patient with expressive aphasia will pause, substitute words, or stop mid-sentence. A patient without aphasia will give you a sentence. “¿Cómo está?” followed by “bien” tells you nothing about aphasia.

3. The seizure witness history where “se cayó y no me respondía” is not a history

The witness history in an acute seizure determines the seizure type (focal vs. generalized), the duration (short focal seizures vs. prolonged events approaching status), the postictal course (which differentiates seizure from syncope and guides next steps), and the first-seizure status (which changes both the workup and the driving conversation). All four of these clinical decisions depend on a structured timeline that the family member almost never arrives with spontaneously.

“Se cayó y no me respondía” is a description of two data points: a fall and a loss of responsiveness. It is not a seizure history. The nurse who documents “patient had witnessed seizure per family, duration unknown” and moves on has preserved the patient’s privacy but not their safety.

The structured timeline that converts a frightened family member’s four-word summary into an actionable history:

Step 1 — Establish where they were and what was happening.

“Vamos a ir juntos por lo que pasó, paso a paso. ¿Dónde estaban cuando empezó?”

(Let’s go through what happened together, step by step. Where were you when it started?)

This is not a clinical question. It is a grounding question. A family member who is re-experiencing the event as they describe it will answer “no sé” to abstract questions but will answer a spatial question. Once they are in the room where it happened, the rest of the timeline follows.

Step 2 — What did you notice first?

“¿Qué fue lo primero que notó— antes de que se cayera?”

(What was the first thing you noticed—before he fell?)

The “before he fell” phrase specifically targets the focal onset you will miss if you only ask about the fall. The family member who says “estaba comiendo y de repente se quedó viendo sin parpadear” (he was eating and suddenly he was staring without blinking) has just told you the seizure began with a focal absence—which changes the type, the workup, and the likely etiology from the generalized tonic-clonic you would have assumed from “se cayó.”

Step 3 — What did the movement look like?

“¿Le tembló todo el cuerpo o solo una parte? ¿Cuál parte primero?”

(Did his whole body shake or just part of it? Which part first?)

Focal clonic onset (one arm, one leg, one side of face) before generalization is a lateralizing sign. A family member who says “primero el brazo derecho, y después todo” (first the right arm, then everything) has just given you a focal-to-bilateral tonic-clonic seizure and a lateralization clue in a single sentence.

Step 4 — How long did it last?

“¿Cuánto tiempo duró el temblor? ¿Fue más de un minuto? ¿Más de dos? ¿Más de cinco?”

(How long did the shaking last? Was it more than a minute? More than two? More than five?)

Witnesses consistently underestimate seizure duration by 50–70%. The event that felt like thirty seconds was more often two minutes. Use minute anchors rather than asking for an estimate: “¿fue más de un minuto?” produces a yes/no response that is more reliable than “¿cuánto duró?” which produces a number that reflects the witness’s distress, not the elapsed time.

Step 5 — Was he breathing? Was he blue?

“¿Respiraba mientras le temblaba? ¿Se puso morado o azul en la cara o en los labios?”

(Was he breathing while he was shaking? Did he turn purple or blue in the face or lips?)

Cyanosis during the ictal phase is a red flag for prolonged convulsion, inadequate ventilation, or an event that was not a typical seizure. The family member who says “sí, se puso un poco morado” has given you a specific finding. Document it as reported.

Step 6 — What happened in the first five to ten minutes after?

“¿Qué pasó cuando paró? ¿Abrió los ojos? ¿Sabía dónde estaba? ¿Hablaba?”

(What happened when it stopped? Did he open his eyes? Did he know where he was? Was he talking?)

The postictal course is what most separates seizure from syncope. A patient who regained full consciousness and orientation within thirty seconds did not have a generalized tonic-clonic seizure. A patient who was confused for fifteen minutes after the motor activity stopped had a postictal period that needs to be in the chart. The family member who says “se quedó dormido como veinte minutos y después no sabía quién era yo” (he fell asleep for about twenty minutes and then did not know who I was) has described a significant postictal period. “Duration unknown, postictal course unknown” is not an equivalent.

Additional language: the altered mental status screen and the “ya me pasó” patient

Two additional conversations recur frequently in neurology nursing in Spanish and depend on specific language choices.

Orientation questions that do not invite social accommodation

Standard orientation questions—“¿qué día es hoy?”, “¿dónde estamos?”— are accurate tests when the patient cannot read the nurse’s face. They are unreliable when the patient is socially oriented toward giving the correct answer.

The patient with mild altered mental status will watch the nurse’s expression after each answer and adjust. A hesitation followed by “¿martes?” is not a correct answer; it is a guess shaped by the nurse’s response to the initial incorrect answer.

Three techniques that reduce social accommodation:

Ask with a neutral face and no vocal inflection. If your expression is readable—and most clinicians’ faces are readable to distressed patients—write the question on a whiteboard and cover the text until the patient answers. This is not theatrical; it is removing a confound.

Start with year before day.

“¿En qué año estamos?”

(What year is it?)

Year is the hardest orientation item to confabulate quickly. A patient who answers 2019 is more disoriented than the day-of-week question would have revealed. A patient who answers correctly has likely maintained the year-level orientation that usually persists longest in organic confusion.

Ask for specifics after a general answer. The patient who answers “en un hospital” to “¿dónde estamos?” may be oriented to situation but not to place. Follow with:

“¿En qué ciudad?”

(In what city?)

Document the specific answer rather than “oriented ×3.” The patient who says “en un hospital en Los Ángeles” is oriented ×2 (situation and general place) but not ×3. That distinction matters in a deteriorating patient over twelve hours.

The “ya me pasó” TIA patient

The transient ischemic attack patient presents a specific problem: the symptoms are gone by the time of clinical contact. The patient’s subjective experience is that nothing is wrong anymore. The nurse’s clinical task is to convey urgency to a patient who currently feels well.

The failure mode:

“Ya me pasó. ¿Por qué tengo que quedarme?”

(It passed. Why do I have to stay?)

The answer that closes this conversation before it works: “Porque el médico lo ordenó.” (Because the doctor ordered it.) The patient who does not understand why they are being detained will continue to ask to leave, escalate to the physician, or leave AMA.

The explanation that gives the patient a reason to stay:

“Lo que pasó es que su cerebro no recibió suficiente sangre por un momento corto, y los síntomas se fueron solos. Eso es bueno. El problema es que cuando esto pasa, el riesgo de que pase algo más serio — un derrame de verdad, donde los síntomas no se van — es más alto ahora mismo que en cualquier otro momento de su vida. El peligro no pasó porque los síntomas pasaron. El peligro sube durante las próximas horas. Por eso estamos aquí ahora.”

(What happened is that your brain did not get enough blood for a short moment, and the symptoms went away on their own. That is good. The problem is that when this happens, the risk of something more serious — a real stroke, where the symptoms do not go away — is higher right now than at any other moment in your life. The danger did not pass when the symptoms passed. The danger rises over the next few hours. That is why we are here right now.)

The phrase “el peligro sube durante las próximas horas” reframes the patient’s recovery as a window of heightened risk, not as a resolution. Patients who hear this framing consistently report it is the first time they understood why they were being kept. Patients who hear “es una precaución” (it is a precaution) understand it as optional.

For the patient who still wants to leave:

“Tiene derecho a irse. Antes de que tome esa decisión, ¿puedo decirle una cosa que me importa que sepa?”

(You have the right to leave. Before you make that decision, may I tell you one thing I want you to know?)

Naming the patient’s right first removes the adversarial dynamic. Most patients who are told they have the right to leave will pause long enough to hear the one thing. The one thing: the 48-hour risk period, what a stroke looks like (sudden arm weakness, face droop, speech change), and what to do if it happens (911 immediately, not to call the clinic first). Give that information whether the patient stays or goes. A patient who leaves AMA and recognizes a stroke two hours later and calls 911 immediately is a better outcome than a patient who leaves AMA and calls the clinic six hours later because they thought symptoms would pass again.

Frequently asked questions

ClinicaLingo builds 10-minute clinical-Spanish scenarios for working US nurses, EMTs, PAs and front-desk staff. Try 29 free scenarios — no login required — or download the free 50-phrase PDF for tomorrow’s shift. Also see: Spanish for neurology nurses — phrase reference, Stroke assessment in Spanish, Spanish for ICU nurses, Spanish for telemetry nurses, Spanish for critical access hospital nurses, When the interpreter is on hold, Family as witness, not interpreter, and the full blog index.