Migraine Diary: What to Track, What to Skip, and 12 Habits That Make It Stick
A migraine diary is the most useful thing you can bring to a neurologist — and the thing most people quietly abandon after three weeks. Here is what is actually worth recording, what to leave out, and the habits that keep it going through the months that matter.
Why a diary is worth the trouble
Memory is the problem. After a bad month you remember the worst attack, not how many there were. Ask someone how many headache days they had last month and the answer is usually wrong in one direction or the other — sometimes badly. A diary replaces an impression with a count.
Three decisions hang on that count, and none of them can be answered from memory:
- What you actually have. The usual clinical line between episodic and chronic migraine is headache on 15 or more days a month for more than three months, with migraine features on at least 8 of those days. That is a counting question.
- Whether a preventive is worth starting — and, once started, whether it is working. Preventives are judged on frequency and severity over months, against a baseline you only have if you wrote it down before you started.
- Whether your acute medication has become part of the problem. This is the one almost nobody spots unaided, because each individual dose feels justified.
A diary is also the difference between a ten-minute appointment spent reconstructing the last three months and a ten-minute appointment spent deciding what to do next.
The minimum useful entry
Five fields. If you record nothing else, record these:
| When it started | A timestamp. Not "Tuesday-ish". |
|---|---|
| How bad it got | Peak pain on a 1–10 scale — and a second reading later if it changed a lot. |
| What you took, and when | Drug, dose, time. Every dose, not just the first. |
| Whether it helped | Helped / partly / no effect. Rated per dose. |
| When it ended | The single most-skipped field, and the one that gives you duration. |
An entry that takes ten seconds gets made during an attack. An entry with fourteen fields gets skipped — and a skipped entry is worse than a rough one, because gaps in a diary read as good days to anyone looking at it later, including you.
What to skip
More tracking is not better tracking. A few things are worth deliberately leaving out:
- Trigger verdicts. Write down that you had red wine; do not write down that red wine caused it. Craving and appetite changes are part of the prodrome, so the "trigger" you reached for may have been an early symptom of an attack that had already started.
- Daily mood and stress scores, unless a clinician specifically asked for them. They are laborious, unreliable after the fact, and rarely change a decision.
- Detailed food logs. Weeks of meal entries very seldom produce a clean signal, and they turn eating into surveillance. If you have a specific suspect, test that one thing.
- Anything that makes you check the diary between attacks. A diary that pulls you into scanning yourself for symptoms is doing harm — that hypervigilance is itself a known amplifier of chronic pain.
The test for any field is simple: could this change a decision my doctor or I will make? If not, it is a chore with no payoff.
Twelve habits that make a diary work
The failure mode is never the first week. It is week five, during a bad stretch, when logging is hardest and the data matters most. These are the practical adjustments that survive that week.
1. Start the log first, describe it later
The only thing that has to happen during an attack is marking that it began. Symptoms, triggers and notes can all be filled in afterwards, when you can look at a screen again. Any diary that demands detail at onset will be abandoned exactly when it is most needed.
2. Make the entry a tap, not a form
Whatever tool you use, get the "an attack is happening" action down to one action from a cold start — home screen, widget, or a voice command to Siri. If it costs more than that, you will do it from memory later, and later is where the accuracy goes.
3. Log the end, not just the start
Duration matters clinically — an untreated migraine attack typically runs 4 to 72 hours, and where yours falls in that range says something about what is working. Ending the attack in the diary also keeps a two-day migraine from being counted as two separate ones.
4. Record every dose separately, and rate each one
Long attacks mean second doses, rescue meds, and things you took at 3am. Logging "sumatriptan" once loses all of that. Rating each dose — helped, partly, no effect — is what eventually answers "does this drug actually work for me, or does it work only when I take it early?"
5. Count medication days, not pills
Medication-overuse headache is defined in days per month, not tablets: broadly, 10 or more days a month for triptans, ergots, opioids and combination painkillers, or 15 or more for simple analgesics, sustained over three months. Two tablets on one day is one day. A running monthly day count is the single most valuable number a diary can show you, and it is nearly impossible to keep in your head.
6. Note the prodrome, not just the pain
Yawning, neck stiffness, mood shifts, food cravings, unusual thirst or frequent urination can arrive hours to a couple of days before the headache. Recording them does two things: it teaches you your own early-warning pattern, and — because acute treatment generally works better taken early — it widens the window in which treatment is worth taking.
7. Log candidate triggers as flags, never as causes
Record the exposure; let the pattern argue for itself across dozens of attacks. Most single-attack trigger attributions do not survive being counted, mostly because of the prodrome problem above and because you only notice the wine on the nights you got a migraine.
8. Track sleep as regularity, not hours
Irregular timing — the weekend lie-in, the short night before a flight — is the part of sleep most often implicated in migraine. Bedtime and wake time are more informative than a total, and if your phone or watch already records sleep, let it fill this in rather than typing it.
9. Keep a short "what actually helped" list
Dark room, ice at the back of the neck, lying down, caffeine, sleep, a specific drug taken within the first half hour. Most people have three or four things that reliably help and rediscover them from scratch every time. Written down, that list becomes a plan you can follow at pain level 8, when you are in no condition to reason.
10. Let context be captured passively
Barometric pressure, sleep, cycle day: if these can be attached automatically at onset, they cost you nothing and you get months of them. If they need typing, you will log them for two weeks. Passive fields are the only ones that survive a bad month intact.
11. Make quiet days visible
An empty diary is ambiguous — it means either "no attacks" or "gave up". A month view that shows every day, attack or not, removes the ambiguity, and it is the one view worth showing a doctor without explanation.
12. Attach logging to something you already do
Backfilling yesterday's attack works best pinned to an existing habit — brushing your teeth, charging your phone. Not a daily reminder notification: those get dismissed on the days you are worst, which is precisely when the entry mattered.
A diary built around those habits
Throb is a migraine diary designed for exactly this: one tap to start an attack, a dark-room attack screen you can use at pain level 8, backfill afterwards instead of during, a pain-over-time curve, per-dose medication logging with a running monthly acute-med day count, and a one-page doctor report PDF with MIDAS. Barometric pressure and sleep are attached automatically at onset.
It is 100% offline. No account, no cloud, no tracking, no ads — every attack stays on your iPhone. The diary and the current-month report are free; a one-time purchase unlocks full history, any-range reports and export. Never a subscription.
iPhone · English, German, French, Italian, Spanish and Japanese · CSV and JSON export, so your data leaves with you.
Reading your own data without fooling yourself
Once you have a few months of entries, the temptation is to find the answer. A few guardrails:
- Co-occurrence is not cause. "Seven of my last ten attacks followed a poor night's sleep" is a real observation. It is not proof, and the arrow may point the other way.
- Count the non-events too. The number that matters is not how often a suspect preceded an attack, but how often it appeared and nothing happened. Almost nobody records that half on their own — which is why the honest version of this analysis usually deflates a favourite theory.
- Small numbers lie confidently. Four attacks is an anecdote. Patterns worth acting on need dozens of entries and, usually, months.
- Test one thing at a time, deliberately, and preferably with your clinician. Cutting five suspected triggers at once tells you nothing except that your life got smaller.
Turning a diary into a better appointment
Neurology appointments are short. Arrive with numbers rather than narrative:
- Headache days per month for the last three months, and how many were migrainous.
- Acute medication days per month per drug — the number from habit 5.
- Typical and worst duration, and typical peak severity.
- What you have tried, at what dose, for how long, and what it did.
- A disability score. MIDAS is five questions about days lost to headache at work, at home and socially over the past three months. It takes a minute, it is widely recognised, and it converts "it's been rough" into something that supports a treatment decision.
One page. A calendar of the last three months, the counts above, and a medication list will do more than a spreadsheet of every entry — which nobody will read during a consultation.
When a headache is not a diary entry but a reason to be seen
Some headaches need urgent assessment rather than logging. Seek medical care promptly for:
- A headache that reaches maximum intensity within seconds to a minute ("thunderclap"), or the worst headache of your life
- Headache with fever, a stiff neck, a rash, or confusion
- Headache with weakness, numbness, trouble speaking, or vision loss — especially if it is new, one-sided, or does not resolve as an aura would
- Headache after a head injury
- A new or clearly changed headache pattern after age 50, during pregnancy, or if you are immunosuppressed or have cancer
- Headache that is consistently worse when lying down, straining or coughing
This list is not exhaustive. If a headache frightens you or feels unlike your usual ones, get it assessed.
Frequently asked questions
What should I write in a migraine diary?
At minimum: when the attack started, how bad it got, what you took and when, whether it helped, and when it ended. Symptoms, possible triggers and where the pain sat are useful additions — but fill them in after the attack, not during it.
How long do I need to keep one before it is useful?
About three months. That is the window the main clinical questions are framed in — chronic versus episodic migraine, medication-overuse thresholds, and MIDAS are all defined over roughly that period. A single month is enough to start a conversation, not to settle one.
Does a migraine diary have to be an app?
No. Paper works, and a paper diary you actually keep beats an app you do not. What an app adds is automatic timing, running counts you cannot hold in your head (acute-med days especially), passive context like barometric pressure and sleep, and a report you can hand over without transcribing anything.
Can a diary tell me my triggers?
Sometimes, and less often than people hope. A diary is reliable at counting attacks, medication use and duration. Trigger identification is genuinely hard: the prodrome can masquerade as a trigger, and you notice suspects mainly on the days you got an attack. Treat any pattern as a hypothesis to test, not a finding.
What is MIDAS?
The Migraine Disability Assessment: five questions about days lost to headache at work or school, at home, and in social or leisure activities over the previous three months. It produces a score and a grade, and it is one of the standard ways clinicians gauge impact.
How many headache days count as chronic migraine?
The usual definition is headache on 15 or more days per month for more than three months, with migraine features on at least 8 of those days. Diagnosis is a clinician's call — but the count that the call rests on has to come from a diary.