Head-to-head comparisons
Every cell in a comparison comes from the same graded evidence as the individual intervention pages — we do not write a separate verdict for a comparison page and we do not soften a grade to make a table look balanced.
Where one side has been graded for an outcome and the other has not, the table says so. That asymmetry is usually the most useful thing on the page.
0 comparisons
No comparisons published yet
A comparison is only published when both interventions are graded and a reviewer has written the editorial read. Until then, compare the individual pages directly.
What a comparison cannot tell you
Two interventions graded for the same outcome have almost never been tested against each other. Comparing their grades compares the strength of two separate bodies of evidence, not the size of two effects — a B graded on three good trials can easily have a smaller real effect than a C graded on one. Where a genuine head-to-head trial exists, we say so on the page and lean on it.