Skip to content

Learn / Correlation vs Causation

Correlation vs causation

After does not necessarily mean because of

A symptom appearing after a medicine begins is important information. It does not, by itself, establish that the medicine caused the symptom. This is the site’s foundational distinction.

8 minute readEvidence last reviewed: 1 September 2026

Happened after does not automatically mean happened because of. If this site has a house rule, that is it. The rule is not a way of talking people out of their experience. It is a way of stopping experience from being stolen by a conclusion that the data cannot carry.

Questions that preserve sequence

  • Temporal association — did it happen afterwards?
  • Dose response — did it change when the dose changed?
  • Dechallenge — did it improve when exposure stopped or reduced?
  • Rechallenge — did it recur if exposure resumed?
  • Baseline — was it already present?
  • Confounding — what else changed?

None of these automatically proves causality in an individual. Together they make a better conversation than “the tablet did this” or “the tablet did nothing”.

Confounders that get forgotten on purpose

  • Another medicine, or a reduction of another medicine.
  • Sleep.
  • Stress, illness, substance use, hormonal change.
  • The underlying condition returning, or being noticed properly for the first time.
  • Previous symptom history.
  • Environmental events.
  • Chance.

Words that are not interchangeable

  • Report — someone described an event.
  • Proportion — a count over a known group.
  • Incidence — new events in a defined population over time.
  • Association — a statistical relationship.
  • Causality — a claim that one thing produced another.
Claim
  1. 1 · Statement

    An adverse-event report to the TGA documents a suspected relationship. It does not prove that a medicine caused an event, and DAEN counts are not incidence rates.

  2. 2 · Classification

    Established. This is how spontaneous pharmacovigilance systems are defined by the regulator.

  3. 3 · Papers

  4. 4 · Claim limitation

    Under-reporting, missing denominators, and confounding are structural.

  5. Last reviewed

    1 September 2026

We are not asking science to be thrown out. We are asking for the evidence loop to include the life it is about.

Walk through a sequence

Sources used on this page

Paradoxical.life provides evidence and self-documentation tools, not medical advice. If you have concerns about your health, medication or mental state, speak with a qualified medical practitioner.