CGRP changed migraine. Your tracker didn't notice.

The shot changed everything. The app did not.

Something specific happened in 2018. The FDA approved erenumab (Aimovig) in May, then fremanezumab (Ajovy) in September, then galcanezumab (Emgality) in September of the same year. Three CGRP monoclonal antibodies in five months. For the first time, there was a class of preventive drugs built specifically for migraine, not repurposed from cardiology or psychiatry, not tolerated as a lesser evil, but designed from the mechanism up.

By 2026, CGRP-targeting therapies (the injectable biologics plus the oral gepants used both acutely and preventively) are mainstream. Neurologists prescribe them. Insurers cover them, reluctantly and after step therapy, but they cover them. Patients are on monthly or quarterly injections as a baseline, sometimes stacked with a gepant like atogepant or rimegepant for breakthrough attacks.

The apps those patients use to track their migraine cycles were, most of them, built before any of this existed. Their data models reflect a simpler picture: you have a headache, you rate the pain, you log what you took, you move on. That picture was incomplete in 2016. In 2026 it has become genuinely misleading for anyone on a CGRP preventive.

The four-gap framework: where CGRP-era tracking breaks

The failure isn't one thing. It's structural, four places where the tracker's assumed model diverges from what life on a CGRP preventive actually looks like.

Gap 1: Attacks look different now, and most trackers can't capture that. CGRP blockade doesn't eliminate migraine; it attenuates it. A lot of patients on effective preventive therapy report attacks that don't reach full pain threshold, prodrome and aura fire, the postdrome materializes, but the headache phase is abbreviated or absent. Standard trackers treat pain score as the anchor. If pain is low or zero, many patients don't even open the app. The attack goes unlogged, and the patient's cycle becomes invisible.

Gap 2: The unit of observation is still the attack, not the cycle. CGRP preventives work across a monthly cycle; the injection date, the titration phase, the trough at week three before the next dose, these shape the whole pattern of frequency and severity. But trackers record events, not cycles. There's no place to mark the injection date as a phase-anchor, no way to visualize frequency relative to the dosing window. A patient trying to assess whether their biologic is working has to do that arithmetic themselves, usually in a notes app.

Gap 3: The postdrome is treated as a footnote, not a data point. This was always a gap. CGRP therapy makes it more obvious because attenuated attacks shift the symptom center of gravity. When the headache phase is mild, the postdrome, the 24-to-72-hour tail of fatigue, brain fog, residual photophobia, neck stiffness, mood dip, can actually dominate the subjective experience of the attack. Trackers that end the attack log when pain resolves are recording the least relevant part of the cycle for a significant share of CGRP-era patients.

Gap 4: The data export doesn't serve the clinical conversation. Neurologists managing patients on CGRP biologics need to see frequency, severity trend, and functional impact across a 3-to-6 month window to decide whether to continue, adjust dose, or switch agents. What most trackers export is a list of timestamped pain scores. The clinical picture a neurologist needs to have a real conversation about titration or switching is not in that export.

What the numbers say about the CGRP moment

The scale of the shift is worth grounding. The American Migraine Foundation estimates that roughly 39 million Americans live with migraine. CGRP-targeting therapies have moved fast: a CGRP injectable biologic treatments are, globally, used by only 2% of drug-treated migraine patients, per IQVIA Patient Link calculations (as of April 2022). ([source](https://www.iqvia.com/blogs/2022/04/migraine-market-continues-exponential-growth)) figure for patients currently on a CGRP-targeting therapy, whether injectable biologic or oral gepant, represents a meaningful slice of that population.

On the clinical side, the landmark Phase 3 trials for the three approved anti-CGRP monoclonals all reported roughly 50 percent of patients achieving at least a 50 percent reduction in monthly migraine days. That responder definition, the 50/50 threshold the field uses, tells you something important: these drugs work, substantially, for about half the people who take them. Which means a large group is on a preventive, experiencing partial response, and trying to assess whether that partial response is enough to justify the cost and the injection burden. That is exactly the person who most needs a tracker that can show them their cycle clearly. And it's the person most let down by a tracker that records pain events without the context of a dosing schedule.

What a typical week looks like now, and what the tracker misses

Take a real scenario. A patient has been on galcanezumab for seven months. She gives herself the 120mg injection on the first of the month. By day 10 or so she's in her clearest window, fewest attacks, lowest severity if one comes. By day 21, she notices her neck is tighter, her sleep is lighter, she's photosensitive in bright offices. She isn't in a headache yet, but she's premonitory. Day 25 she gets a moderate attack, pain peaks at maybe a 5, she takes her gepant, pain resolves in three hours. But then the next day she can't think clearly. The day after, she's exhausted and her neck stiffness is back. Day 28, the fog lifts. Day 32, she injects again.

Her tracker, if she opened it, would show: one pain event, severity 5, 3-hour duration, medication logged. That's the record of a 7-day cycle that included prodrome, a compressed headache phase, a 48-hour postdrome, and a 3-week pattern shaped by the biologic's pharmacokinetics.

The next time she sees her neurologist, the conversation is: 'How are you doing on the galcanezumab?' She says 'Better, I think, but I still have bad weeks at the end of the month.' The neurologist looks at the app export, a list of pain events, and can't see the dosing-cycle pattern. They can't see that the postdrome is where she's losing two days a month. The conversation is less useful than it should be, and the data that could make it useful was never captured.

Why trackers haven't updated, and why that's a design choice, not just lag

The most charitable reading is that apps built before 2018 simply haven't rebuilt their core data model. A tracker designed around pain-event logging has that shape baked into its database, its exports, its UI logic. Rebuilding around a cycle-and-phase model takes re-architecture, not just a new field.

But there's a less charitable reading. The most widely used migraine apps are built around engagement loops: daily check-ins, streak protection, habit formation. Those loops work best when the app is the center of the patient's health behavior, not a clinical documentation tool. A CGRP-era tracker built for clinical utility would look different. It would prioritize the injection date as a first-class object. It would extend the attack log into the postdrome. It would generate an export that a neurologist could read in 90 seconds. None of those features are good for daily-engagement metrics.

The result is that the category has optimized for retention, opening the app every day, over utility, giving the patient a complete picture of their cycle and giving their neurologist something actionable. Those are different products. Most apps built the first one.

What changes if you fix the model

A tracker built for the CGRP era has a different core assumption: the thing worth capturing is the full symptom arc, anchored to the treatment cycle, not a sequence of pain events.

That means the injection date goes in as a phase marker. The prodrome, aura, headache, and postdrome all have their own status fields, so an attenuated attack with no headache but two days of brain fog still gets recorded. The export isn't a pain log; it's a frequency-and-severity timeline that makes sense in a clinical conversation about whether a biologic is working.

And it means the postdrome gets the same design attention as the headache phase. Logging when you can't see straight, can't concentrate, can't tolerate light, that's a real UX problem. Aura Mode, the ability to start a log with one tap when full interaction is impossible, is the kind of feature that only makes sense once you've accepted that the postdrome is where a meaningful share of migraine burden actually lives.

The patient who benefits isn't the newly diagnosed person deciding whether to try a triptan. It's the person seven months into a biologic, trying to figure out whether the postdrome is getting shorter, whether the end-of-cycle vulnerability is real, whether they have something concrete to show their neurologist at the next appointment. That person exists in large numbers right now. Most of them are using a tracker that wasn't designed for them.

What Postdrome tracks, and why we built it this way

Postdrome was designed from the postdrome backward. Not 'log a headache, optionally add aftermath', but 'the attack isn't over when the pain stops, so the log shouldn't be either.'

The injection date is a first-class field. The attack log extends through prodrome, aura, headache, and postdrome phases, each with its own symptom set, because residual photophobia in the postdrome is different from photophobia at headache peak, and recording both separately is what makes the timeline readable to a neurologist. Aura Mode is a one-tap entry for when full interaction isn't possible. The monthly view shows frequency relative to your dosing window, not just dots on a calendar.

The export is built for clinical conversations. Not a PDF of pain scores, but a structured timeline a neurologist can scan before the appointment and actually use.

We built this because the people we built it for are on a treatment that changed migraine medicine in 2018, and they're still using tools that don't know that.