Plain English

The words on this sitemean specific things.

Every number here comes wrapped in vocabulary borrowed from epidemiology. None of it is decoration — each term draws a hard line around what the number can and cannot tell you. Here is what each one actually means.

Hazard ratio

also written HR

How much faster something happens in one group than in another, over the same stretch of time. 1.0 means no difference between the groups. 1.30 means the outcome arrived 30% faster in the group being compared; 0.70 means 30% slower.

It is a comparison between two groups of people. It is not your odds, and it is not a percentage of anything that belongs to you.

The limit

A hazard ratio never says which direction the causation runs. People who walk more may live longer because walking helps, or they may walk more because they were already well.

Confidence interval

also written 95% CI

The range the true value plausibly sits in, given the data. Written after the estimate: HR 0.70 (0.59–0.83) means the best guess is 0.70 and the data are consistent with anything from 0.59 to 0.83.

Narrow interval, more certainty. Wide interval, less. A study reporting a wide interval is telling you it is not sure, and that is worth reading as carefully as the estimate itself.

The limit

If the interval crosses 1.0, the study did not find a difference it could tell apart from chance — whatever the headline number looks like.

All-cause mortality

also written ACM

Death from anything at all. Not death from the disease being studied — death, full stop.

It is the hardest endpoint to argue with, because nothing can be quietly traded away. A drug that prevents heart attacks but causes cancer looks excellent on one endpoint and completely flat on this one. Most numbers on this site use it.

The limit

One row here does not. Blood pressure is measured against heart and stroke death specifically, which is a narrower claim, and the calculator says so on that row.

Cohort study

also written Prospective cohort

Take a large group of people, measure something about them, then wait years and record what happens. Nobody is assigned to do anything; you observe what people were already doing anyway.

Almost everything on this site is built from cohorts. They are how you get to study 60,000 people for seven years, which no trial will ever pay for.

The limit

A cohort can show you that two things move together. It cannot show you that one caused the other. That gap is the reason the evidence grades on this site stop where they do.

Meta-analysis

Pooling the results of many separate studies into one combined estimate. More people, tighter confidence interval, and a number that does not depend on any single research group having got it right.

When a row here cites something like "48 studies," that is a meta-analysis.

The limit

It inherits every bias the original studies carried in with them. A meta-analysis of weak studies is a precise weak answer, not a strong one.

Biomarker

Something measurable about your body that stands in for something you actually care about. Grip strength is not the point — what it says about whole-body strength and how well you are ageing is the point.

A good biomarker is cheap to measure and closely tied to the thing it represents. The eight on this site were chosen on both counts.

The limit

A marker can move without the thing it stands for moving at all. Training your grip specifically has not been shown to buy you the years that having a strong grip predicts.

Index

A single number built by averaging several others together, used for ranking rather than for measuring. Your MarkerScore is one. 50 is a typical person on all eight. The scale goes to 100; that is a cap, not a test score.

An index is useful for deciding what to work on: it puts eight incomparable things on one scale so you can see which of yours is furthest from the reference. That is the whole job it does.

The limit

An index, not a risk estimate. 50 is a typical person on all eight. The scale goes to 100; that is a cap, not a test score. This is a weighted average of single-marker associations from published studies. The weights are a judgment call, and the combined number has never been validated against health outcomes — no study has measured all 8 markers on the same people. Use it to decide what to work on, not to estimate your risk.

Evidence grade

A letter on every claim saying how hard that specific claim has been tested. It grades the claim, never the metric — the same marker can carry an A on one endpoint and an X on another.

Someone ran the experiment on that outcome. A trial that finds nothing is still A — the letter is how hard it was tested, not whether the result was flattering.

Large groups followed for years — not trials. Dose-response in a cohort, or a genetic natural experiment.

The numbers moved together in a group of people. That is not the same as one causing the other.

A plausible mechanism. No human outcome study for this claim.

We looked. This protocol has not been tested against that outcome.

Where the numbers live

Every claim shows its work.

Each of the eight markers opens the study its number came from, the grade that claim earned, and what argues against it.

See the metrics

Free · No sign-up · Nothing to install