About MarkerScore

MarkerScore is a calculator that turns eight everyday health metrics — things like steps, sleep, VO₂ max, HbA1c, and blood pressure — into a single directional score, built entirely from hazard ratios published in peer-reviewed cohort studies and trials. Research estimates trace to cited sources; constructed model choices and estimated measurement costs are labeled.

It is not medical advice, a diagnosis, or a prediction about you as an individual. It is not a medical device, and it has not been evaluated or approved by the FDA or any other regulatory body. It does not compute "years of life" added or lost — you will not find that framing anywhere on this site, because the underlying studies do not support translating a hazard ratio into a life-span figure for a person. The score is a population-level index: it tells you how your entered values compare to hazard ratios observed in large groups of people, not what will happen to you. See the full medical disclaimer for the complete list of what this tool cannot do.

Each of the 8 scored metrics has its own hazard-ratio curve, built from published cohort studies and meta-analyses via monotonic interpolation between literature-derived anchor points — so risk changes smoothly rather than jumping at arbitrary thresholds. The 8 metrics are grouped into three physiological domains (physical capacity, cardiometabolic health, and lifestyle); each domain collapses to a single number via a weighted geometric mean, and the three domains combine the same way into the composite. Because the weights average rather than accumulate, the composite ranks a profile against a population reference, but it does not — and is not claimed to — compound risk the way a jointly-fitted clinical model would.

Every curve, every anchor point, and every domain weight is documented, with its primary source, on the Methodology & Sources page, grouped by domain and metric. How the score is calculated is on that page. It is not behind a login.

Every evidence-backed claim on the site — not just the 8 scored metrics, but the extended library under More markers — carries a letter grade from A to X. The grade describes the design of the study behind the claim, not our confidence in the result and not whether the finding was flattering. A trial that finds no effect is still graded A if it was a well-run trial on that endpoint.

• Readouts, not exposures. Most of the 8 scored metrics — VO2 max, resting heart rate, HbA1c, blood pressure — describe what your body is doing now; they are not the behaviors (smoking, alcohol, diet quality, body weight) that mortality studies typically rank as the largest modifiable risks. A person can smoke, drink heavily, and eat poorly and still score well on all 8. That is a scope choice, not a finding that those exposures don't matter.
• An age-invariant reference profile. Your score is measured against one fixed reference profile, a 45-year-old, at every age. The ranking shown next to it is different: it compares you with people of the same sex in the same ten-year age band, as a range. Neither is a clinical result.
• Constructed curves. Between published anchor points, hazard-ratio curves are interpolated, not directly measured. The anchors come from cited studies; the curve connecting them is a modeling choice, documented as such.
• An index, not a combined risk estimate. The composite score is a weighted average of log hazard ratios. It ranks profiles against a population median; it does not compound risk the way a jointly-fitted multivariable clinical model would, because no single cohort has measured all 8 markers on the same people. It is never expressed as a percentage change in risk.
• Self-reported inputs. The score is only as accurate as the values you enter. Self-reported steps, sleep, and diet carry real measurement error compared to clinical measurement.

For the full list of what the tool cannot do, see the medical disclaimer.

MarkerScore carries no affiliate links, no sponsorships, and no advertising. The calculator and the methodology do not require an account. The site loads no ad network and no tracking pixel. Vercel Analytics and Umami count page views without cookies or cross-site tracking. Vercel Speed Insights measures web performance, also without cookies. How the score is calculated should stay visible. No commercial incentive sits behind a scored number on the page.

Questions about the methodology, a citation that looks wrong, or anything else about the site can be sent to hello@markerscore.com.

Corrections are welcome and are the fastest way to improve the site. If a cited study does not say what the page claims it says, that is a bug — send it over and it gets fixed or the claim comes down.