Your headache diary was built for a drug trial, not for you.

Day three, and the app wants a yes or no

It's the third morning after the attack. The pain left sometime Monday night. What stayed: a dull ache at the base of the skull when you turn your head, light off the bathroom tile that still feels a notch too loud, and a word you reach for mid-sentence and can't find.

You open your tracking app. It asks one question.

Did you have a migraine today?

Yes inflates your count. No erases the last 48 hours. So you pick one, or you close the app, and either way the record gets worse. Anyone who has tracked through a chronic stretch knows this screen. And plenty of them decide the fault is theirs.

The protocol inheritance

The headache diary on your phone is a clinical trial instrument wearing consumer clothes. We call this the protocol inheritance. The format's core unit, a discrete attack with a start, a peak and an end, was standardized to answer a regulator's question: does this drug beat placebo? Consumer apps took that unit and kept it, mostly without asking whether a patient's question is the same one.

It isn't. A trial needs a countable event so it can compare two arms cleanly. A person living with near-daily symptoms needs to see what the days actually looked like, including the hours a protocol was never built to score.

Why the friction explanation misses

The standard diagnosis for abandoned migraine diaries is friction. Too many taps, too many fields, forgotten reminders. So the fixes are faster logging, home-screen widgets, a nudge at 9 p.m. Each of them makes the binary quicker to answer.

None of them changes the question.

Episodic migraine has edges. An attack starts, peaks, resolves, and the days between are mostly yours. Chronic migraine, by the ICHD-3 definition, means headache on 15 or more days a month for more than three months. Past that line the edges blur into a state, and a tool that only recognizes events starts recording noise. People quit because the diary is describing someone else's disease. Calling that an adherence problem blames the patient for the instrument.

What the trial guidance actually measures

The lineage is public. The International Headache Society published its first guidelines for controlled trials of drugs in migraine in 1991 and has revised them since, with separate guidance for acute and preventive studies. The US FDA's 2018 guidance for industry on developing acute migraine treatments recommends two co-primary endpoints measured two hours after dosing: freedom from pain, and freedom from the patient's most bothersome symptom. Preventive trials center on change in monthly migraine days from baseline, with the share of patients reaching a 50% reduction as a standard secondary measure.

Every one of those endpoints needs a diary. And every one is satisfied by exactly the fields consumer apps ship. When did it start? How bad was it? Did you take something? Was today a migraine day? Put the trial endpoint next to the app field and the match is close to one-to-one, right down to what neither of them asks about.

One chronic week, scored two ways

Take an illustrative week for someone near the chronic threshold. Monday: a full attack, treated with a triptan, pain gone by evening. Tuesday and Wednesday: no headache worth the name, but fog, neck stiffness and residual photophobia heavy enough to cancel plans. Thursday: a low-grade ache from noon that never builds. Friday: clear until 9 p.m., when the yawning starts. Saturday: attack. Sunday: postdrome again.

A binary diary scores that week as two migraine days. Three, if Thursday gets the benefit of the doubt. The person lived seven days shaped by the disease, and the record a neurologist sees suggests the preventive is doing fine.

The failure is built into the format. A yes/no diary offers two answers, and this week needs more. Log Tuesday as a yes and it claims a headache that never happened. Log it as a no and the cancelled plans disappear. Either way the record loses something, and the closer someone gets to daily symptoms, the more it loses.

A paper tracker doesn't have that limit. More columns, a few levels of severity, and room for the in-between days can hold the whole week: the attack, the postdrome, the ache that never builds. An app that only asks "migraine today?" can't hold that, so the record that reaches the neurologist is the thinner one.

Keep the count, change where it comes from

None of this argues for dropping the monthly number. Headache days still decide a chronic diagnosis, and prior-authorization paperwork for preventive treatment often asks for them. The count matters.

What should change is its source. A count typed in as a nightly yes or no is a judgment call made by a tired person at 11 p.m. A count derived from a timestamped record of symptoms is something you can check, recompute against a different definition, and defend when an insurer or a new neurologist asks how you got it.

Research has already moved in this direction for its own questions. Studies of the migraine postdrome, the phase after the pain ends, have needed to capture what happens across the whole attack, not just the headache itself. A single yes or no per day can't show that. The science moved past the binary when it needed to see the after-phase. Patient tools mostly stayed put.

Logging without a declared endpoint

Postdrome starts from the other assumption. You log symptoms as they happen, each with a timestamp, and nothing in the app asks you to open or close an attack. Fog at 2 p.m. on day two is an entry. Neck stiffness on day three is an entry. A monthly headache-day count becomes something you can work out from the timeline and show a clinician, instead of the only question the app ever asked.

The rest follows from designing for the person and not the protocol. Aura Mode for the hours when looking at a screen is its own problem. Your data stays on your device, and the export is yours to hand to whoever needs it. Lifetime pricing is stated up front.

What we're tracking next

Whether the next revisions of trial guidance start treating between-attack burden as something worth measuring. Instruments for it already exist in the research literature, such as the Migraine Interictal Burden Scale (MIBS-4). If regulators begin counting the days between attacks, consumer apps will copy that too, the same way they copied the attack.

We'd rather not wait for permission to record the week you actually had.