How to Interview a Dizzy Patient

A teaching guide for reading vertigo, one question at a time

By Will Sheppy, L.Ac.

Start here: the one idea everything hangs on

Before I teach you a single question, learn this, because it changes how you'll hear everything the patient says:

A beginner asks questions to fill out a chart. An experienced clinician asks questions to knock possibilities off a list. Every question you ask should do one of four jobs — characterize the dizziness, test a diagnosis, isolate a system, or plan the treatment — and every answer should either rule something in or rule something out.

Picture a funnel. A patient walks in and the funnel is wide: it could be the ear, the eyes, the neck, the blood pressure, the nerves, crystals, migraine, anxiety. Your questions narrow it. By the end, one or two possibilities are left standing, and that's your diagnosis — not because you matched a name to a list of symptoms, but because you eliminated everything else.

Keep the funnel in your head. Now let's fill it.

The method: six stages, in order

I'll walk you through the six stages the way I actually use them. For each one I'll give you the goal, the questions, and the logic — and I'll show you the answers from a real patient so you can watch the funnel narrow.

(Our patient: dizzy since a fall where she hit her head on concrete twice. Came in already diagnosed with Ménière's by an ENT. Watch what we do with that.)

Stage 1 — Characterize it: "What kind of dizzy?"

Goal: Find out what the patient actually means, because "dizzy" is four different words wearing one coat.

Ask:

• "What does your dizziness feel like?"
• "Do you see the room spinning?"
• "How long does it last?"

The logic: Teach your ear to sort the answer into one of four buckets, because each points somewhere different:

And watch the clock. Duration is one of the most powerful sorting tools you have — memorize this and you're halfway home:

• Seconds → position/neck
• Minutes → crystals (BPPV)
• Hours → Ménière's or migraine

Teaching point: Never let a patient's word choice do your thinking. Make them describe the sensation until you can place it in a bucket.

Stage 2 — Test the label: "Does the diagnosis they were handed actually fit?"

Goal: When a patient arrives with a diagnosis, your job is not to accept it. Your job is to test whether their pattern matches it.

Ask (the specific fingerprints of Ménière's):

• "Do you have ringing in the ear (tinnitus)?"
• "Any fullness or pressure in the ear?"
• "When an attack hits, how long and how severe is it?"

The logic: Every diagnosis has a fingerprint — a small set of features that should be present if the label is real. Ménière's is a fluid problem in the inner ear, so its fingerprint is: true spinning vertigo, tinnitus, ear fullness, often hearing loss, and long, severe attacks lasting hours. Check for each print. Misses matter as much as hits.

Teaching point: A diagnosis is a hypothesis someone else wrote down. Treat it as a claim to be tested, not a fact to be inherited.

Stage 3 — Isolate the system: "Which leg of the stool is failing?"

Goal: Find which part of the balance system is the problem by removing variables one at a time — like an experiment.

First, the anatomy you're testing. Balance stands on three legs:

1. The inner ear (motion sensor)
2. The eyes (they and the ear are in constant conversation — when the ear says "we're turning," the eyes track the opposite way to hold your gaze steady)
3. Proprioception (your body's sense of position, mostly from the neck and joints)

Ask (and notice the design of these questions):

• "Sitting still — are you dizzy?" (removes the legs and standing)
• "Now, sitting still, move your head around — does it jump?" (with legs removed, this tests only ear/eye/neck)
• "Does looking up make it worse?" (a targeted provocation of the neck)

The logic: This is the most important teaching move in the whole interview: change one variable at a time. If sitting is fine but sitting-and-moving-the-head is not, you've just proven the legs aren't the problem and the upper circuit is. You're not guessing — you're running a controlled test on a living system.

Teaching point: When you can't tell where a problem lives, subtract one input and see if the problem survives. What remains is your answer.

Stage 4 — Eliminate the usual suspects, by name

Goal: Even with the region identified, you must rule out the other common causes individually. Each is its own hypothesis with its own test.

Ask, suspect by suspect:

• Crystals (BPPV): "Is turning one way worse than the other? Does lying down and getting up set it off?"
  Logic: BPPV is positional and one-sided, but crystals take a minute to five minutes to tumble through the canal. If the episode is 30 seconds, it's too short to be crystals.
• Blood pressure (orthostatic): "Did the dizziness start when you began your blood-pressure medication?"
  Logic: Here the tool is timeline. She's been on the medication for years; the dizziness started after the fall. The timing clears it.
• Migraine: "Do you get headaches?"
  Logic: Vestibular migraine is common and sneaky. No headaches drops it down the list.
• Anxiety / nervous system: "When you're dizzy, do you sweat or does your heart race?"
  Logic: Autonomic dizziness announces itself with sweating and a racing heart. Neither present → not panic-driven.
• The eyes: "Any blurriness? Does driving in a car trigger it?"
  Logic: A visual-vestibular mismatch shows up in the eyes. Neither present → sharpens the focus back onto the neck.

Teaching point: Don't skip a suspect just because you've got a hunch. Each "no" you collect makes your final "yes" stronger.

Stage 5 — Find the origin: "When did it start, and why?"

Goal: The story of when and why often names the culprit outright.

Ask:

• "Is it constant, or does it come and go?"
• "What was happening when it started?"

The logic: Constant vs. episodic reshapes the whole picture, and the origin event is often the whole answer. A symptom that begins the instant something happened usually belongs to that something.

A bonus lesson — the canary. Her family noticed that fluids and food snap her out of the worst spells. Don't misread that as the cause. Once a balance system is already weak, it becomes the canary in the coal mine — the first thing to complain when anything else (dehydration, low blood sugar, fatigue) goes sideways. The trigger isn't the disease; it's the weakest link showing strain first.

Teaching point: Always ask what was happening at the very first episode. The origin is frequently the diagnosis.

Stage 6 — Confirm and plan: "Prove you're right before you commit."

Goal: These last questions aren't diagnostic — they set up treatment and give you a test that confirms your theory.

Ask:

• "Have you had bodywork before? How did you tolerate it?" (so you know how to position and pace them)
• "When your chiropractor adjusts your neck, does it ever change the dizziness — trigger it or relieve it?"

The logic: That second question is a confirmation test. If working the neck changes the dizziness in either direction, the neck "talks to" the symptom and you're on the right system. Then set the patient's expectation honestly: "When I treat the neck today, it may briefly stir the dizziness — and that's good news, it means we found the right target. If nothing changes, that's information too, and we'll adjust."

Teaching point: The best clinicians build a test into their treatment. Treat, reassess, and let the body tell you whether you were right.

The whole method on one line

Run that funnel honestly and the diagnosis falls out on its own. For our patient: a brief, position-triggered, neck-driven coordination problem that started with a fall — not Ménière's.

The rules of thumb, so you can do this without me

Memorize these seven and you can walk into any dizzy patient's room:

1. Don't collect symptoms — eliminate possibilities. Every question narrows the funnel.
2. Make them describe the sensation until you can name the bucket. Never trust the word "dizzy."
3. Let duration sort the field. Seconds, minutes, hours each point somewhere different.
4. Treat any handed-down diagnosis as a hypothesis and check it against its fingerprint.
5. Change one variable at a time. Subtract an input; see if the problem survives.
6. Ask what was happening at the very first episode. The origin is often the answer.
7. Build a confirmation test into your treatment. Treat, reassess, let the body answer.

Now you try

Next dizzy patient you see, don't reach for a name. Reach for the funnel. Ask "what does it feel like," start subtracting variables, and watch how few questions it takes before the diagnosis is standing there alone. That's the whole art: not knowing more names, but asking better questions.