BEFORE THE APPOINTMENT
Ask your neurologist five things. What makes this the diagnosis, and what would change their mind? What do your headache-day and medication-day counts mean? Is any of it medication overuse? What is the preventive half of the plan? And what counts as working, by when?
Most people leave a specialist appointment having answered questions rather than asked any. These are the ones that change what happens next.
Ask these five
- “What specifically makes you think this is the diagnosis, and what would change your mind?” A good answer names features from your history. An answer that only describes a scan is describing a picture, not you.
- “What are my two numbers, and what do they mean for what I am eligible for?” Headache days and medication days set the thresholds for several treatments. If nobody has counted with you, that is the gap.
- “Is any part of this medication overuse, and if so what is the plan for unwinding it?” Nothing else can be judged fairly until this is answered.
- “What is the preventive half of the plan?” If you only leave with something to take during an attack, you have half a plan.
- “What counts as this working, by when, and what happens if it does not?” A stated endpoint in days per month, with a review date, is the difference between a plan and a prescription.
The question to ask about anything you have already tried
Not “what else is there” but “was that a complete trial?” A preventive stopped after two weeks, or never titrated (raised step by step) to an effective dose, is an incomplete trial. It gets recorded as a failure. That closes off a whole class you could still use. Bring doses and durations, and ask directly whether each was adequate.
What to bring so the answers are worth having
Two weeks of diary, the list of what you have tried with doses and durations, any imaging reports, and a current medication list including anything over the counter. The full list.
And one thing to say out loud
If your headaches have been dismissed before, say so at the start rather than hoping it will not happen again. People who have been disbelieved tend to understate — leaving out how often it happens, or what actually helps — and the edited history is the one that fails to produce a diagnosis. Oddly, being explicit about the history of not being believed is one of the most useful things you can do for an accurate record.
Frequently asked questions
What if I forget to ask?
Write the five on the back of the diary. Nobody minds you reading from a card. It is the single most reliable way to leave with a plan rather than a prescription.
What a plan’s stated endpoint and review points look like over the first three months is set out in how long a headache workup takes.
Is it rude to ask what would change their mind?
No. It is a normal clinical question and a good clinician will answer it directly. An answer that cannot be given is itself informative.
Should I ask about a scan?
Ask what a scan would change. If the honest answer is nothing, that is the correct answer — what imaging settles.
What if I disagree with the plan?
Say so in the room, and ask what the alternative would be. A second opinion is also reasonable — when it is worth getting one.
Take these with you
Leave with an endpoint, not just a prescription
A plan states what success looks like and when it is reviewed. That is the thing to insist on.
12174 Natural Bridge Rd, Suite 304
St. Louis, MO 63044
Next to DePaul Hospital, off the I‑270 and I‑70 junction, west of the airport.
Sources
- Steiner TJ et al. Aids to management of headache disorders in primary care (2nd edition) : on behalf of the European Headache Federation and Lifting The Burden: the Global Campaign against Headache. The journal of headache and pain, 2019. PubMed 31113373
- Tram J et al. Interdisciplinary Headache Management. Physical medicine and rehabilitation clinics of North America, 2025. PubMed 41167859
- Pohl H. Red flags in headache care. Headache, 2022. PubMed 35167120
- Diener HC et al. Drugs for Migraine Prophylaxis. Deutsches Arzteblatt international, 2026. PubMed 41572863