What to measure so guidance can be tailored to you. Health evidence speaks in strata — reference classes, not individuals — so tailored advice needs a few numbers to place you in the right one. This is the input list: the minimal, readily-capturable set that does most of that placing, ranked by decision-impact × capturability. A metric earns a spot only if knowing it changes a stratum or a decision — not because it can be measured.

This is a capture list, not a diagnosis. Reference ranges, interpretation, and acting on a number are clinician acts. Naming what to measure is in scope; what it means for you is not.

Tier 1 — free or trivial, and they do most of the placing

  • Age and sex — define the reference class and are the first inputs to any cardiovascular-risk estimate (SCORE2 Baseline Risk and the ESC Treatment Thresholds).
  • Smoking status — the single largest modifiable exposure, and a risk-estimate input. Free to ask.
  • Blood pressure — the cardiometabolic lever with the best-proven primary-prevention benefit (Blood Pressure Lowering and Cardiovascular Events); a home cuff suffices.
  • Resting heart rate — a cheap co-marker and an input to a no-test fitness estimate (below).
  • Waist circumference (plus BMI) — visceral adiposity is the dominant metabolic lever and a MASLD criterion (Fatty Liver MASLD and Weight Loss); waist beats BMI for visceral fat, but BMI is universal and free, so record both. A tape measure is the whole apparatus.

Tier 2 — one standard blood draw

  • HbA1c — places the dysglycaemia / metabolic-syndrome stratum (Carbohydrate Restriction and Type 2 Diabetes Remission); preferred over fasting glucose (no fast, less day-to-day noise).
  • Lipid panel → read non-HDL. Non-HDL is a cheap, standard apoB proxy — it captures all the atherogenic, apoB-bearing particles — and for the metabolically impaired it beats LDL-C, which under-counts the particle burden when small, dense LDL predominates (LDL ApoB and Cumulative Exposure). Read non-HDL, not LDL-C, as the headline here.
  • ALT — the cheapest MASLD screen (Fatty Liver MASLD and Weight Loss); standard, and this stratum leans MASLD.

Tier 3 — add only where it significantly moves a decision

  • Lp(a) — an independent, genetically-set apoB particle (LDL ApoB and Cumulative Exposure); measure once in a lifetime, and decision-changing for the ~1 in 5 who carry a high level. High yield for a one-off test.
  • ApoB — the direct particle count, better than any proxy, but not yet on every standard panel. Take it as the upgrade over non-HDL where available, not a baseline requirement.
  • VO2max / cardiorespiratory fitness — among the strongest mortality predictors (Cardiorespiratory Fitness and Mortality), but it needs a test — so the non-exercise estimate (from age, resting heart rate and activity, all already in Tier 1) is the population-available route (Measuring and Raising Cardiorespiratory Fitness).
  • Grip strength — a cheap dynamometer, and among the strongest mortality predictors (Muscle- Strengthening Activity and Mortality); not yet a standard measure but trivial to add.

What NOT to capture (it doesn’t move the needle enough)

  • Body-fat % (DEXA/BIA) — waist is the cheaper, population-available visceral proxy and carries most of the decision value; the scans aren’t standard. Waist covers it.
  • hsCRP — it moves an inflammatory marker but is weakly tied to decisions, and the wiki’s own worked case shows the marker can fall while outcomes don’t (Does Weight Loss Reduce Cardiovascular Events: CRP dropped ~42% with no event benefit). Optional, not baseline.
  • Fasting glucose — largely redundant with HbA1c. Skip as a separate line.
  • LDL-C as the headline lipid — kept on the panel, but read non-HDL / apoB instead here (LDL-C under-counts the particle burden). Not dropped — de-emphasised.

The honest limits

  • This grades what changes a decision, not what improves your health. And the wiki holds no head-to-head added-value study (apoB over non-HDL, VO2max over the estimate, grip over nothing) — the tiering rests on predictor strength, not a reclassification trial.
  • Capture, not interpret or prescribe: this names what to measure; ranges, diagnosis and action are a clinician’s.
  • A starting default, tuned per person: which metrics matter most shifts by stratum (a lean 25-year-old and the drifting-median adult need different emphases). The tiers are the default, not a fixed prescription.