38 conditions · 214 graded rows
Conditions
What each condition has actually been studied with, and which compounds were checked and produced nothing.
Last updated
The compounds that earned nothing are on the page too
306 compound-and-condition checks produced 214 graded rows. The other 89 are named on the condition they were checked against, so a list of five is never mistaken for the only five anyone considered.
Groupings
A curated hierarchy, not an inferred one. A child rolls up to its parent for counting and keeps its own identity — dyspareunia and osteoporosis are both menopause, and a reader looking for one does not want the other.
- Heart failure — 2 narrower indications, and 9 graded rows of its own
- Hyperglycaemia — 1 narrower indication, and 10 graded rows of its own
- Hyperpigmentation — 1 narrower indication, and 3 graded rows of its own
- Menopause — 8 narrower indications, and 5 graded rows of its own
- Pain — 2 narrower indications, and 4 graded rows of its own
- Skin ageing — 3 narrower indications, and 5 graded rows of its own
- Weight — 5 narrower indications, and no rows of its own
Rung and directness are two different judgements
The rung rates the study design and says nothing about what was measured. Directness rates what was measured and says nothing about how well. A randomised human trial that measured a mitochondrial marker is a human RCT and it is not evidence about the condition, so every row carries both.
- Direct outcome
- The study measured the condition itself, or a symptom a patient would report.
direct_outcome - Surrogate marker
- It measured a biological marker accepted as standing in for the condition, not the condition.
surrogate_marker - Mechanistic only
- It measured something upstream. The link to the condition is an argument, not a result — these rows are recorded and labelled, never presented as evidence for the condition.
mechanistic_only
Approval is per condition, and per population
An approval next door is not an approval here. A compound approved for premenopausal women is
not_approved on a menopause row, with the adjacent approval explained in the note —
collapsing that distinction is how a record turns into a recommendation.
- Approved for this condition
- A national regulator approved it for this condition. The note names the regulator.
approved - Off-label use
- Approved for something else, and prescribed for this.
off_label - In registered trials
- In registered clinical trials for this condition, and not approved.
investigational - Not approved for this condition
- No approval for this condition anywhere. This is the common case.
not_approved - Approval not established
- We could not establish an approval status. The note says what was checked.
unknown
How a condition gets a page
Conditions are discovered from the corpus rather than enumerated from medicine: each record is asked what it has been studied for, the answers are collapsed into canonical names, and a name earns a page once at least two compounds link to it. Below that the link is recorded and the page is not built.