How to Tell If Your Migraine Medication Is Actually Working

Judge it across at least three attacks, not one. For each, record when the pain started, when you took the drug, how bad the pain was at that moment, whether you were pain-free two hours later, and whether it stayed away for the rest of the day without another dose.

That last pair is the endpoint drug trials use, and it is far stricter than the question most people ask themselves. “Did it help?” almost always gets a yes, because pain that peaks tends to fall afterwards whether or not you took anything.

Pain-free at two hours, and still pain-free at 24

Acute migraine drugs are tested against two things: the share of attacks that are pain-free two hours after dosing, and the share that stay pain-free through 24 hours without needing a rescue dose. The second number is always lower than the first, and it is the one that matters for how your day actually went.

Both come from the International Headache Society’s guidelines for migraine drug trials, which set pain freedom at two hours as the primary endpoint and sustained pain freedom — pain-free at two hours, no rescue medication, no relapse through 24 hours — as the stricter secondary one. The guidelines moved to pain freedom rather than “pain relief” for a specific reason: people treated to mild pain did not report it as a success, so the older, easier endpoint was measuring something patients did not want.

Adopt the same bar. “Took the edge off” is not what these drugs are supposed to do — a triptan that reliably leaves you at a four out of ten is a triptan that is not working well enough, and that is worth saying out loud to a prescriber rather than accepting.

Two hours is the checkpoint because it is when the effect should have landed. If you are still in full pain at the two-hour mark, that attack is a failure for that drug at that dose taken at that time — which is three separate variables, and the next section is about the one people get wrong.

The timing confound that makes a good drug look useless

Triptans work substantially better taken while the pain is still mild than once an attack is fully established. Wait until you are at an eight, and the same tablet that would have worked at a three may do very little. The trials of early intervention with oral triptans reviewed by Dowson, Mathew and Pascual point the same way across several drugs: dosing while pain is still mild produces markedly higher pain-free rates than dosing once it is severe.

This wrecks most self-assessment. Someone tries sumatriptan on four attacks: twice they caught it early and it worked, twice they waited it out hoping it would pass and it didn’t. Averaged together the drug looks fifty-fifty and unreliable. Separated by when it was taken, it looks like a drug that works when taken early.

So record the pain level at the moment you dose, not just the peak. Without that number you cannot tell a drug failure from a timing failure, and the two have completely different fixes.

There is a real tension here worth naming: taking something early on every twinge is how you end up using acute medication on too many days a month, which brings its own problem. Treating early and treating rarely pull against each other. That tension is a genuine conversation to have with a doctor, not something to resolve from a blog post.

Why one attack proves nothing

Migraine attacks vary enormously on their own. Some resolve in three hours untreated; some run for two days. If you take a drug during a short attack, the drug looks brilliant. Take the identical drug during a long one and it looks worthless.

There is also a statistical trap. You take medication when the pain is at its worst — that is what prompts you. Pain at its worst has nowhere to go but down. Some of the improvement you attribute to the tablet would have happened anyway, and there is no way to tell how much from a single instance.

The defence is repetition. Across enough attacks, the noise averages out and a real effect separates from the drift. Three attacks is the usual minimum before judging an acute treatment; more is better, and it is why this needs a record rather than a memory.

“It came back” is a different result from “it didn’t work”

If you were pain-free at two hours and the headache returned that evening, the drug worked and then wore off. That is recurrence, and it points somewhere entirely different from a drug that never touched the attack — often at how long the particular drug stays in your system relative to how long your attacks run.

Score it as its own outcome. Lumping “never worked” and “worked then came back” together as “didn’t work” hides the single most actionable pattern in your own data. Recurrence has its own page: why your migraine comes back after the medication works.

What to write down

Six fields per attack, and no more, because a longer list gets abandoned:

• Time the pain started
• Time you took something — the gap between these two is the timing variable
• Pain level when you dosed, on whatever scale you use consistently
• What you took, by name
• Pain-free at two hours? Yes or no, not a description
• Did it come back within the day? And did you need a second dose

Six fields, thirty seconds, and after a handful of attacks you have something no amount of reflection produces.

Migraine Journal records what you took per attack and whether it helped, then shows the helped-of-rated share across every dose of that medication — which is the multi-attack aggregate this whole page is arguing for. Be aware it asks the coarser question: helped or didn’t, rather than pain-free at two hours. If you want the stricter endpoint, the notes field is the place for it. There is also a free printable diary if paper suits you better.

What to do with the answer

If a drug fails across several attacks that you treated early, at a full dose, you have something concrete: “I have tried this on five attacks, took it within thirty minutes each time, and was still in pain at two hours on four of them.”

That sentence is actionable in a way that “I don’t think it really works” is not, and it belongs on the page you hand over — what to bring to a neurologist appointment covers where the treatment history sits alongside the four figures. There are several triptans and they are not interchangeable — people who fail on one often respond to another — and there are non-triptan acute options now. Frequent failures are also one of the standard prompts for discussing preventive treatment rather than only better rescue. Preventives are then judged on a different clock entirely — how long before a migraine preventive works is counted in weeks, and a fair trial has a defined minimum, which is why so many are abandoned a fortnight early.

Measuring properly also decides what you become eligible for. The frequency thresholds on the newer preventives are written in documented days — how many migraine days it takes to qualify for Botox is the clearest example, and undercounting is the common way people sit just below a line they actually meet.

The point of measuring properly is to arrive at that conversation with evidence instead of an impression. The other tracking guides are built around the same multi-attack record.

---

This is background reading, not medical advice. Dose, timing and which drug to try are decisions for whoever prescribes for you. If your attacks have changed in character, or a headache comes on suddenly and severely, that needs a doctor rather than a diary.