Is a trigger the same as a cause?

What does a diary actually observe?

A diary records that two things happened near each other in time: a short night and a headache the next day, chocolate and a migraine, a pressure drop and joint pain. That is an association. It says nothing yet about direction or mechanism. In 1965 Austin Bradford Hill set out the questions an association has to answer before anyone should call it causal, among them strength, consistency, timing and whether a dose makes a difference (Hill, 1965). A personal diary can speak to a few of those. It cannot speak to all of them, and it cannot run the experiment that would settle it.

Why do so many triggers turn out to be something else?

Prodrome. Migraine is the classic case. Attacks often begin hours before the pain with cravings, yawning, mood changes and light sensitivity. A person who eats chocolate during that phase and gets a headache afterwards will record chocolate as the trigger, when it was the attack that produced the craving. Surveys find that most people with migraine report triggers (Kelman, 2007); careful diary studies find the picture far less tidy.

Reverse timing. When stress was tracked daily and compared with headache onset, the link was not with high stress but with the drop in stress afterwards, the “let-down” effect (Lipton et al., 2014). A diary that only asked “were you stressed today?” would have recorded the wrong half of the story.

Coincidence that repeats. If you track ten factors against one symptom for a month, several will line up by chance. Pick the strongest and it will look convincing. This is the multiple-comparisons problem, and the usual remedy is to demand more of a link the more candidates you tested (Benjamini & Hochberg, 1995).

Confounding. Late nights go with alcohol, skipped meals and irregular medication. A link with any one of them may belong to another.

How should you word what you see?

Wording matters because it shapes behaviour. People give up foods, sleep, exercise and social lives on the strength of triggers that were never tested, and the cost of that is real.

When is a trigger worth acting on?

Ask four questions of your own record:

1. Does it hold on enough days? Not three late nights but twenty, with a comparable number of early ones.
2. Does it hold at a sensible timing? Same day or next day, and not only when you shift the window to make it fit.
3. Does it survive the number of things you compared? One link out of forty candidates needs to be much stronger than one out of two.
4. Does it repeat? Next month, with no change in how you log, is it still there?

If the answer is yes to all four, you have something worth raising with a clinician and, if it is harmless to test, worth testing deliberately. That is still not proof. It is a good reason.

How does Longnote word associations?

Longnote never writes “cause” or “trigger” as a conclusion. A link is reported with the counts on both sides, the number of overlapping days, an evidence word, and a standing: observation, hint, early sign or pattern. Weather and other factor cards end with the line associations, not causes. Patterns are only promoted after a false-discovery check across every comparison the app made, and Highlights show at most one early sign at a time, because in testing a second one was wrong more often than it was right.