← All categories

Evidence

Topic 11

The evidence ladder

Why it matters: Different kinds of studies answer different questions, and not all results carry the same weight.

Think of evidence as a ladder: Cells → Animals → Human trials → Regulatory approval. Each step can add information, but approval is a decision, not a kind of experiment.

A result in cells is an early clue, a bit like testing a car part before testing the whole car. Studies do not always happen in this exact order, and a higher step does not guarantee a perfect answer.

  1. step 1Cell (test-tube) studies
  2. step 2Animal studies
  3. step 3Small human trials
  4. step 4Large human trials
  5. step 5Regulator approval

Topic 12

Evaluating a study

Why it matters: A study's design helps determine how much confidence to place in its findings.

Think of a study as a test of a mystery. If only a few people take part, an unusual result may be due to chance rather than a real pattern.

A fair comparison matters too. Researchers may keep participants and investigators from knowing who received what, and other teams can repeat the work to see whether the finding holds up.

Topic 13

Dose and response

Why it matters: More isn't always better: changing an amount can change both effects and risks.

Think of a volume knob: turning it up changes the sound, but louder is not always better. Likewise, changing the amount of a substance can change what happens.

There may be a range where a useful effect is more likely and harm is less likely. That range can differ from person to person, and more of a substance may increase harm without adding benefit.

amount →
Solid: wanted effect. Dashed: unwanted effect. Shaded: the "window" where benefit outweighs harm. Concept only — no real numbers.

Topic 14

The translation gap

Why it matters: A promising result in mice does not necessarily predict what will happen in people.

Imagine testing a bridge design using a small model. The model can reveal useful clues, but it cannot show everything that will happen to a full-size bridge in the real world.

Mice and humans have some biology in common, but they also differ. A mouse study may model only part of a human condition, so its results need careful testing before anyone knows whether they apply to people.