Methodology
Published study estimates are cited to their sources: studies that follow people over time, trials, or reviews that combine studies. The calculator's curves also contain constructed points, flat regions and assumptions. Values between published points are model values, not findings reported by a study. Below is how we grade evidence, how the model works, then the bibliography grouped by domain.
Page last reviewed 2026-10-08. Changes to the scoring model are dated in the changelog. Weights are below.
Want to see how everyday habits feed these markers? The habit map traces each connection to its studies.
Four walks
The first tranche. Past the bends, more of the same buys less. These are not a plan and not a save.
Walk to the step bend
Ding et al. 2025, 14 prospective cohorts, all-cause mortality: about 7,000 steps a day versus 2,000, point estimate HR 0.53. That comparison is not the on-screen step target. Paluch et al. 2022, 15 cohorts, all-cause mortality: the displayed target is 8,000 steps a day under age 60 and 6,000 from age 60. Age only selects that target. It does not change the step curve.
Two short lift days
None is the reference. Two days a week, counted here as 30 minutes a day, is where the line flattens. The row can only raise the score, and it overlaps steps and VO₂ max. Momma 2022.
Fiber floor from food
The displayed fiber target is IOM/NASEM adequate intake: a gram floor of 38, 30, 25, or 21 g by sex and age. Age and sex only select that floor. They do not change the fiber hazard curve, which is absolute grams per day from Reynolds et al. 2019, all-cause mortality. Reynolds’s categorical peak at 25–29 g is a different number.
Sleep window around seven hours
Both sides of about seven hours are associated with higher all-cause mortality. The long side is steeper. Yin 2017.
What your score means
Your score compares your eight numbers with a typical 45-year-old's. 50 means you match them on all eight. Above 50, your numbers look more like those of people with lower death rates in large studies; below 50, more like those with higher death rates. It's a comparison, not a prediction.
Not yet in the score: aerobic minutes, sitting time.
Because almost nobody is typical on all eight, a 50 already beats almost 2 in 3 people.
How much each metric counts
The eight scored metrics, each with its study and grade, are on The 8 metrics. Markers that do not move your number are on More markers.
The composite is a weighted index of associations from large observational studies. It has never been validated as a combined clinical predictor — no cohort has measured all eight markers on the same people and tracked what happened to them — and association in these studies is not proof that changing a marker changes your risk by the same amount.
Physical Capacity · 44% of the composite
Cardiorespiratory fitness, movement, & strength training
10 cited sources — see the bibliography below
7 cited sources — see the bibliography below
8 cited sources — see the bibliography below
10 cited sources — see the bibliography below
Cardiometabolic · 34% of the composite
Vascular & glycemic control
15 cited sources — see the bibliography below
8 cited sources — see the bibliography below
Lifestyle · 22% of the composite
Sleep & nutrition
6 cited sources — see the bibliography below
6 cited sources — see the bibliography below
What happens if you leave a metric blank
Once you start entering values, any of the eight you leave untouched counts as typical (the median reference) in your score, so it neither helps nor hurts. The calculator still shows a placeholder value on that row, marked as an estimate, but the placeholder does not feed your score. A blank metric keeps its normal share of the composite; it is not dropped, and the other seven are not reweighted to cover for it. A score built from a few entered markers is a real but partial picture, not a confidence-adjusted estimate. Enter all eight for the most complete score.
How we grade evidence
Each chip grades one claim on one endpoint, not a marker as a whole. A marker can carry a B headline on home and A facts on its metric page; those A facts are real and they are not the home number.
- Grade ARCT or meta-analysis of RCTs on that endpoint
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.
- Grade Bprospective cohort with dose-response, or Mendelian randomization
Large groups followed for years — not trials. Dose-response in a cohort, or a genetic natural experiment.
- Grade Ccohort association only
The numbers moved together in a group of people. That is not the same as one causing the other.
- Grade Dmechanistic reasoning only
A plausible mechanism. No human outcome study for this claim.
- Grade Xno evidence for the claimed endpoint
We looked. This protocol has not been tested against that outcome.
Some cards show METHOD instead of a letter — method notes, not claims.
Every Grade A fact on this site is a randomized trial. Most of them did not lower all-cause deaths, and none of the eight scores is built on one. What the trials found.
Per-metric hazard ratios
For measures on a sliding scale, the curve runs between anchor numbers taken from the studies, and the shape between those anchors is our own. It is smooth, so risk changes gradually instead of jumping at cutoffs.
Domain aggregation
Technical notes
Metrics are grouped into three domains; each collapses to a single number via a weighted geometric mean, then the three combine the same way. Weights sum to 1 at both levels, so this is a weighted average of log hazard ratios — one metric moves the result by exactly its own weight share, and nothing is double-counted.
Why the composite is an index, not a risk estimate
Because the weights average rather than accumulate, the composite ranks profiles against the population median (1.00) but does not compound risk across markers. Estimating combined risk honestly would need a jointly-fitted multivariable model, and no cohort has measured all 8 markers on the same people. So the composite is published as an index and is never expressed as a percentage change in risk. Per-metric hazard ratios are different: each is a single-marker claim backed by a single-marker study, which is exactly what the cited papers support.
Score & comparison range
The score stays on its fixed midlife scale: 50 matches the same reference profile on all eight markers at every age. Your modeled ranking compares with the same sex in ages 20-29, 30-39, 40-49, 50-59, 60-69 or 70+. Ages 18-19 use a newly pooled same-sex adults 20+ comparison. Rankings are ranges, rounded outward in five-point steps.
How the citations are checked
Every metric is backed by at least five primary studies, listed in full below. Each quoted number is checked against the sentence in the paper it comes from. If we cannot find that sentence, the number does not ship.
What it is not
This is a directional, educational estimate — not a validated clinical risk model or medical advice.
What the eight do not cover
The scored eight are not a lifestyle score. Most of them are readouts — numbers that describe accumulated physiology — not the exposures mortality studies rank when they ask which actions account for most deaths. A person can smoke, drink heavily, eat poorly, and carry high body weight and still look typical or better on all eight. That is a scope choice, not a finding that those exposures do not matter.
Most of the eight are readouts
VO2 max, resting heart rate, and HbA1c describe what the body is doing now. Steps and fiber are intakes, and strength training is a count of days you did it. Sleep is duration, not regularity or clock time. Blood pressure is the closest exception: it is both a reading and close to the pathway the Global Burden of Disease ranks. That ranking is of disease burden (years lost to death and disability), not a deaths table, and it is a population share, not your personal risk.
Smoking, alcohol, diet quality, and body weight are not in the score
They have no input. Joint lifestyle-score cohorts are built from never smoking, weight in a healthy range, activity, moderate alcohol, and diet quality — the usual core in this literature. Physical activity here is scored only as VO2 max, steps and days of strength training, not as minutes of aerobic exercise. Smoking is the largest single gap against both that core and the GBD ranking; it was considered and declined as a product input. The epidemiology is not in dispute.
The markers outside the eight are on More markers. The bibliography below covers the eight scored metrics.
Physical Capacity sources
VO2 Max
Daily Steps
Resting Heart Rate
Strength Training
Cardiometabolic sources
Blood Pressure
HbA1c
Lifestyle sources
Sleep Duration
Dietary Fiber
Model limitations
We correct calibration issues when we find them, and we disclose them here rather than waiting for someone else to find them first.
Calibration was corrected
Earlier versions anchored several curves’ "no added risk" point (HR = 1.0) at values well below what a typical adult actually does — the daily-steps curve, for example, treated 2,000 steps/day as the neutral baseline, when that is closer to the bottom of the range than the middle. The effect was that ordinary results looked better than they were. Scores are now expressed relative to a defined median reference profile for a 45-year-old US adult, so matching that profile on every marker scores exactly 50 — typical, not a test out of 100. See "Fifty isn’t the middle of the pack" below for why that is a higher bar than it sounds.
What the per-metric hazard ratios mean
Where a study reports a usable number at a named exposure, that anchor is the paper’s figure against that paper’s own comparator. Interiors of the PCHIP curves, and the constructions listed in the next card, are ours — direction from the papers, specific shape not. Recalibration of the composite left those per-metric numbers in place so the published anchors stay traceable. The composite score stays relative to the fixed median reference; the modeled comparison range uses the age and sex group.
More curve shapes are our own construction than three
Technical notes
Grip strength, VO2 max, and HRV are the three we previously named: Leong reports only a per-5kg slope (and, as of 2026-09-10, only on the below-optimal side — the above-optimal decline to a 0.7 floor was removed; the curve now holds flat at 0.8 at and above the 40/25 kg anchor because neither Leong nor Wu 2017 publishes a dose-response for getting stronger past it), Mandsager uses age/sex percentile bins, Jarczok compares quartiles (its HRV curve was rescaled 2026-08-29 to Jarczok’s covariate-adjusted estimate rather than the unadjusted one). Also constructed, and now named: the daily-steps PCHIP interiors (0 / 3k / 4k / 5k / 16k — 2,000 = 1.0, 7,000 = 0.53, 10,000 = 0.52, and 12,000 = 0.45 are all Ding 2025’s own reported figures as of 2026-09-10; Ding models 1000-step increments only up to 12,000/day, "due to scarce data past this point," so 16,000 stays flat at the 12,000 value — flat there is a data-scarcity cap, not a reported plateau); fibre’s curve was rebuilt 2026-09-10 to track Reynolds 2019’s all-cause-mortality dose-response directly — HR = 1.0 at 16 g/day (the median reference intake), moving log-linearly per that paper’s reported RR 0.93 per 8 g/day, over the roughly 5–35 g/day range Reynolds actually modelled (Figure 2a: “linear with no sign of a plateau within the available data”). Outside that range the curve is flat in both directions — below 5 g/day and above 35 g/day alike — because that is where the evidence stops, not because benefit stops or reverses. The age/sex fibre target (IOM/NASEM Adequate Intake, 38/30/25/21 g) is shown as advice only and no longer changes the score — Reynolds’s dose-response has no age or sex term, so scoring the same intake differently by age/sex bucket was never supported by the mortality evidence; sleep duration’s short-side PCHIP anchors (4 h = 1.08, 5 h = 1.04, 6 h = 1.01) are Yin 2017’s own Table 2 nonlinear estimates, same as the long side (8 / 9 / 10 h = 1.04 / 1.15 / 1.32) — changed 2026-09-26 from a steeper linear 1.06-per-hour extrapolation on the short side (1.19 / 1.12 / 1.06); gait speed’s 1.2 m/s recentre for a ~45-year-old (Studenski measured adults 65+); sleep-regularity’s log-linear fill between Windred’s 1.0 and 0.70 quintile endpoints (0.70 is the fully adjusted most-regular HR; 0.52 is the same contrast in the minimal model, not used; interiors 0.91 / 0.84 / 0.76 are ours, not Table 2); social’s three-level 0.85 / 1.0 / 1.25 (Strong / Average / Weak) is constructed around typical = 1.0 — not Wang 2023’s isolated-vs-not 1.32, and 0.85 is not 1/1.32 (isolation, loneliness, and living-alone are three separately pooled constructs); and the blood-pressure combination rule when systolic and diastolic disagree (the high-side shape and magnitude generally — the description previously overclaimed continuous interpolation from the Prospective Studies Collaboration; corrected 2026-09-10). Dental (poor HR 1.15, Wu 2021) and hearing (loss HR 1.13, Tan 2022) are directly-sourced figures, not constructed. Direction comes from the papers; those specific shapes are ours.
Three ranges we could not verify, and did not move
Three curve regions are not traceable to a cited continuous dose-response, and the underlying papers were paywalled (403) rather than confirmed or refuted when we checked — so the constants are disclosed here, not changed. HbA1c above 8% (anchors 8.0 → 1.95, 10.0 → 2.7): no cited source supplies a continuous dose-response reaching 10%; Schöttker’s population is non-diabetic (ceiling ~6.5%), and the other HbA1c citations are open-ended categorical bins — the 2.7 value is constructed. Would be settled by a diabetic-population continuous dose-response reaching 10% HbA1c. VO2 max’s top two knots (male 63/73, female 56/66 mL/kg/min): almost certainly beyond any cohort’s observed range; Singh 2024’s full text could not be opened to confirm its measured ceiling. Resting heart rate’s 40–50 bpm flat band (both pinned to 0.85): Zhang 2016’s own spline reads linear through that low zone (p nonlinearity = 0.1), arguing 40 and 50 should not score identically, but the range could not be confirmed from Zhang’s full text (cmaj.ca returned 403). None of these three constants were changed on inference alone.
Strength training can only raise your score
Strength training (days a week of exercise done to strengthen muscles) is scored at 5%, and the reference is zero days because most US adults do fewer than two a week. That makes it bonus-only: doing none scores as typical, never worse, and it is the one marker that cannot pull your score down. Each day counts as 30 minutes, an assumption with no source. The curve is flat above about an hour a week because the studies disagree there (one large cohort saw no benefit at two hours or more), so more is not scored as better. The evidence is observational and the pooled review grades it very low certainty, so confidence is LOW-MEDIUM. The zero-day reference comes from a secondary summary of a national survey, not a cited paper.
Balance is one clinic cohort, not the 2024 meta
Fail HR 1.84 is Araujo 2022 (CLINIMEX clinic volunteers, 10-second one-legged stance — the same test the calculator asks). Das et al. 2024 pooled 15 cohorts of mixed static-balance tests at HR 1.14 (1.07–1.21), I² 88%; that pool is not one-leg-only. Xie 2023 (CHARLS) is a 10-second semi-tandem stand, a different and easier test (1% failed vs Araujo’s 20%). The score uses Araujo’s figure.
The percentile comes from a simulation, not measured people
Age-band rankings preserve same-person survey measurements where available and draw the remaining markers from separate U.S. survey and exercise-testing data. The range covers proxy and assumed-dependence sensitivity runs plus core survey sampling uncertainty. It is not a fully validated population percentile or a medical result. Some inputs are indirect or older. Ages 70+ combine older survey cells and use the observed 70-79 exercise-testing median with a wider spread because direct 80+ values are missing. Fewer studies include people over 70, so this range is wider.
Fifty isn’t the middle of the pack
Score 50 matches the fixed midlife reference on all eight markers. It does not mean the middle of every age group. Only the modeled comparison changes with age and sex; the score, curves, weights and tier names stay fixed. Saved comparisons keep the age, sex, displayed range and distribution version from that assessment. Legacy saved results retain their old display.
The roadmap projects, it does not promise
Roadmap point gains are computed by re-running this same model with one input changed — so they can never drift away from your score — but they are arithmetic, not outcomes. They assume you hold every other metric where it is, that a change you make is genuine rather than a measurement artefact, and that the association the underlying study observed would apply to you. Observational cohorts show that people with better numbers die less often; they cannot prove that moving your number moves your risk by the amount shown.
Self-reported inputs carry real error
Grip strength, VO2 max, HRV, and sleep regularity in particular depend on the user’s equipment and technique, and are only as accurate as what the user measures and enters.
This is not a medical device
This is a general-wellness and educational tool. It is not a diagnostic device, it does not provide medical advice, and it must not be used to make treatment decisions. Discuss your results with a clinician.
Reference-profile changelog
Technical notes
2026-09-29 (strength training added; sitting reviewed). Strength training joined the composite as an eighth scored marker at 5% of the score. The other seven weights were scaled proportionally (each times 0.95), so their relative ratios are unchanged. Input is days a week (0 to 7), counted as 30 minutes a day; the curve (0, 30, 40, 60 minutes = 1.00, 0.87, 0.83, 0.80, then flat) is built from aerobic-adjusted studies (Momma 2022, PMID 35228201, figures confirmed against the full text; Shailendra 2022, PMID 35599175; Patel 2020, PMID 32762807). The reference is zero days, so the marker can only raise a score, and the reference profile still scores exactly 50. Confidence is LOW-MEDIUM. Sitting time was reviewed and kept out of the score because its effect is largely gone in active people and it cannot be framed as something to do; it is documented, unscored, on the sitting page. Score bands moved to 37/49/56/64/73, and the reference profile now reads ~p62 for men and ~p67 for women, almost 2 in 3.
Older entries
2026-09-29 (note on the percentiles below). The p75-range percentiles quoted in the older entries below are historical. After the 2026-09-26 distribution refresh the median reference profile (raw score 50) read ~p64 for men and ~p69 for women, about 2 in 3; the strength-training change above moved that to ~p62 and ~p67.
2026-09-19 (grip leaves the score). Grip strength was taken out of the composite, which then scored seven markers: steps, sleep, fiber, blood pressure, HbA1c, VO2 max, and resting heart rate. Grip is still published, unscored, on the more metrics page. Domain weights did not change; within Physical Capacity the remaining members are VO2 max 0.47, steps 0.33, and resting heart rate 0.20.
2026-09-10 (steps + grip + BP visualization). Two curve changes and four copy/rendering fixes, all following the same rule as the fibre rebuild: a curve tracks its source only as far as that source actually observed. Steps — Ding et al. 2025 (Lancet Public Health, PMID 40713949) models the all-cause-mortality dose-response in 1000-step increments from 2,000 up to 12,000 steps/day, capped there for data scarcity, not because the paper found a plateau. The curve had pinned 10,000 and 12,000 to an identical HR 0.51, discarding evidence inside Ding's own modelled range. Anchors moved to Ding's own reported figures vs the 2,000-step reference: 10,000 steps 0.51 → 0.52 (Ding: "48% ... in all-cause mortality"), 12,000 steps 0.51 → 0.45 (Ding: "the maximum risk reduction modelled ranged from 12% in cancer incidence to 55% in all-cause mortality"). 16,000 stays flat at the new 0.45 — above 12,000 Ding has no data, so flat is still correct there. Grip — the above-optimal decline (−0.005/kg to a 0.7 floor) is removed; the grip curve now holds flat at 0.8 at and above the 40 kg male / 25 kg female anchor, because neither Leong 2015 nor Wu 2017 publishes a dose-response for grip strength above that anchor, only a per-5kg-lower slope. The below-optimal side is unchanged. Blood pressure — the always-visible target band's start moved from 90 to 114 mmHg systolic (visualization only; the blood-pressure formula is untouched): Wang et al. 2026 (PMID 41411626, UK Biobank, n=80,730) reports the lowest systolic quintile (<114 mmHg) at 16% higher all-cause mortality vs 120–126, so painting all of 90–120 as uniformly "in the zone" contradicted the site's own cited source. Also fixed: the blood-pressure card's description no longer claims "continuous interpolation from the Prospective Studies Collaboration" (false — PSC/Lewington's endpoint is vascular death, not all-cause, and the high-side shape is hand-constructed); the steps habit-picker copy no longer calls 7,000/9,000 steps "at"/"past the plateau" (Ding shows continuing decline to 12,000); resting-heart-rate's 40–50 bpm flat band now has a disclosure symmetric to the existing above-80 one. Three further regions (HbA1c above 8%, VO2 max's top two knots, and the RHR 40–50 flat band itself) were reviewed and found not traceable to a cited dose-response, but were left unchanged — the source papers returned 403 and a mortality curve should not move on an unverified range. See "Three ranges we could not verify, and did not move" above. Re-deriving the distribution after the steps and grip changes moved the score-band cutoffs. Any score computed before this change is not directly comparable to one computed after it for profiles near 10,000–16,000 steps/day or above-optimal grip.
2026-09-10 (fibre curve). The fibre curve was rebuilt to depend only on absolute g/day, tracking Reynolds et al. 2019’s all-cause-mortality dose-response directly instead of the age/sex adequate-intake target. Before this change, a 30-year-old man and a 60-year-old woman each eating 30 g/day of fibre scored different hazard ratios (0.88 vs 0.81) purely because of their age/sex target bucket, with no mortality evidence behind the gap — Reynolds’s dose-response has no age or sex term. The new curve is anchored to HR = 1.0 at 16 g/day (the median reference intake) and moves log-linearly at Reynolds’s reported RR 0.93 per 8 g/day, over the roughly 5–35 g/day range the paper actually modelled (Figure 2a is "linear with no sign of a plateau within the available data" across that range). The curve is now flat outside 5–35 g/day in both directions — it no longer climbs to HR 1.225 for zero intake, and no longer keeps rewarding intake past target+20g the way the prior cap did. Flat means the evidence stops, not that benefit stops. The age/sex fibre target (IOM/NASEM Adequate Intake, 38/30/25/21 g) is unchanged and still shown as the advice line, but it no longer affects the score. Re-deriving the distribution after this change produced identical score-band cutoffs (35/45/52/59/67). Any score computed before this change is not directly comparable to one computed after it for profiles whose fibre intake was near the old age/sex target.
2026-09-10 (labels). The score-band labels were renamed — Below Typical / Typical / Above Typical / Strong / Exceptional / Elite — so the ladder reads monotonically in plain English; the old ladder put "Average" second from the bottom and landed the median person in "Fair." This is a naming change only. It is separate from, and did not cause, the percentile-cut moves logged below — those moved because the underlying model changed (dropping Recovery, then removing the blood-pressure low-end reward), not because the labels did.
2026-09-10. The blood-pressure low-end reward was removed. Below 120/80 the curve had been paying a bonus growing to a −10% floor at 100/68 and below; the best available all-cause-mortality evidence for that range (Wang et al. 2026, UK Biobank, 80,730 adults without traditional cardiovascular risk factors) found the opposite sign — the lowest systolic quintile carried 16% higher risk, not lower. The low side is now flattened to HR 1.0 — no reward, no penalty — below 120/80; the high side is untouched. This is the same change made for the HbA1c low-end reward on 2026-08-29, for the same reason: an unsourced bonus for an extreme value, in a direction the mortality evidence does not support. Re-deriving the distribution after dropping this reward moved the score-band cutoffs from 36/46/53/60/69 to 35/45/52/59/67; some profiles no longer get a below-120/80 bonus, so the whole population's raw scores shift down slightly. The median reference profile still scores exactly 50. Any score computed before this change is not directly comparable to one computed after it.
2026-08-29 (second pass). Separately from the reweight below, the scored stack was cut from 17 metrics to 8 keepers — the composite dropped the Recovery domain (HRV, hearing, dental) and now scores only steps, sleep, fiber, grip, blood pressure, HbA1c, VO2 max, and resting heart rate. Re-deriving the distribution for 8 keepers only moved the score-band cutoffs from 39/47/51/57/63 to 36/46/53/60/69. Any score computed before this change is not directly comparable to one computed after it.
2026-08-29. Four changes were made. Hearing-loss HR moved from an uncited 1.25 to Tan et al. 2022's pooled 1.13. Dental was redefined from a borrowed clinical-periodontitis HR (1.30) to a self-reported-tooth-loss HR of 1.15 (Wu et al. 2021, the only cited cohort using this app's own self-report modality) — the question itself changed from gum-disease status to natural tooth loss; the underlying 'good'/'poor' values and shareable-link encoding did not. HbA1c's low-end reward (HR 0.88 below 5.4%) was removed — Schöttker et al. 2016 and Zhong et al. 2016 found that association was inconsistent, non-significant, or driven by undiagnosed diabetes — so the curve is now flat (HR 1.0) up to 5.6% rather than U-shaped. HRV's 10ms endpoint moved from a constructed 1.40 (the curve never sat at Jarczok's unadjusted 1.56) to Jarczok et al. 2022's covariate-adjusted lowest-quartile estimate (~1.23); the other HRV anchors were pulled toward 1.0 at 33ms with per-point log-space factors of 0.587–0.652, not one uniform scale. HRV's weight within the Recovery domain fell from 0.5 to 0.3 (hearing and dental each rose from 0.25 to 0.35 to renormalize). Re-deriving the distribution after these changes moved the score-band cutoffs from 38/46/51/57/63 to 39/47/51/57/63, and the median reference profile's percentile from ~p75.4-75.8 down to ~p74.6-75.1 (male 74.636, female 75.058). Any score computed before this change is not directly comparable to one computed after it.
2026-07-29. The gait-speed addition below was revised after two problems turned up in the first version. First: the 1.2 m/s reference point is an unsourced judgment call, not a measured figure — Studenski et al. 2011 only measured adults 65+ (pooled mean 0.92 m/s); 1.2 m/s is this app's own recentring for a younger adult, now called out explicitly here, next to the gait-speed slider in the calculator, and in the source note below, rather than being folded silently into the model. Second, the weighting was fixed: the first version funded gaitSpeed's 0.15 weight in the Physical Capacity domain partly by cutting grip strength from 0.15 to 0.10 — grip strength has comparably strong, arguably more broadly-replicated all-cause-mortality evidence, so this read as favoring a newer, more speculative metric over an established one. Grip strength is restored to 0.15; the room for gaitSpeed now comes only from gaitSpeed's own initial allocation (0.15 → 0.10), leaving vo2max, steps, resting heart rate, and balance unchanged throughout. Re-deriving the distribution after this weight change moved the score-band cutoffs from 38/47/52/57/65 to 38/46/51/57/63 (close to the original pre-gait-speed 38/46/52/57/64), and the median reference profile's percentile from ~p73.7-74.1 back up to ~p75.4-75.8. Any score computed before this change is not directly comparable to one computed after it.
2026-07-29. Added usual gait speed (walked over a measured ~4-meter distance) as a 17th metric, in the Physical Capacity domain. Source: Studenski et al. 2011 (JAMA 305(1):50-58, PMID 21205966), a pooled analysis of 9 cohorts covering 34,485 community-dwelling adults 65 and older, one of the most validated all-cause-mortality predictors in geriatrics — age-adjusted HR for death of 0.88 (95% CI 0.87-0.90) per each 0.1 m/s higher gait speed. Because that cohort is older adults (mean age 73.5) and this app models a ~45-year-old reference population, the population median used for scoring (1.2 m/s) is a recentred judgment call rather than the source's own 0.92 m/s pooled mean. Adding a 17th metric shifted the simulated distribution slightly: score-band cutoffs moved from 38/46/52/57/64 to 38/47/52/57/65, and the median reference profile's percentile moved from ~p74.4-75.7 to ~p73.7-74.1. Any score computed before this change is not directly comparable to one computed after it.
2026-07-29. A check of the literature covered every ASSUMED spread parameter behind the simulated population used for percentiles and score bands. Most held up as the best defensible estimate and were left unchanged with stronger citations. One did not: VO2max's population SD was based on an inferred ACSM percentile-band width (6 mL/kg/min for both sexes); the FRIEND registry (Kaminsky et al. 2015, a large directly-measured cardiopulmonary exercise test cohort) reports the real figure as 9.6 (male) / 7.7 (female) — meaningfully wider. That widened the whole simulated distribution: the score-band cutoffs shifted (p80, previously 50, is now 52), and the median reference profile's percentile moved from ~p78.5-80.1 down to ~p74.4-75.7. Any score computed before 2026-07-29 is not directly comparable to one computed after it.
2026-07-27. The median reference profile behind every score was re-anchored on two markers. Daily steps moved from 5,500 steps/day (the same value for both sexes, which had no supporting source) to 8,500 (male) / 7,000 (female), based on NHANES accelerometer data and the All of Us Research Program. Grip strength moved from 42 kg (male) / 26 kg (female) to 41 kg / 29 kg, based on dominant-hand, best-of-trials norms from a NIH Toolbox study — and the model now states explicitly, for the first time, that it uses the dominant-hand convention throughout. VO2 max was reviewed against two disagreeing primary sources and deliberately left unchanged at 38 (male) / 31 (female), with the reasoning for that choice now written up in the model's source comments. Because the reference profile moved, the score-band cutoffs and the percentile shown next to a score also shifted slightly. Any score computed before 2026-07-27 is not directly comparable to one computed after it — re-run your numbers if you want an apples-to-apples comparison.
How much would the score move under different weights?
The weights above are one defensible allocation, not the uniquely correct one. To make that tradeoff inspectable rather than a single judgment call, we recompute the composite for a panel of 22 representative profiles (median-profile and single-metric deviations, both sexes) under four alternative weight sets: equal-weight across domains and members, cardiometabolic-heavy (50/30/20), physical-heavy (55/30/15), and one that swaps the sleep/fiber emphasis inside Lifestyle. The live model uses three domains and eight scored metrics.
The largest single swing observed, across every profile and every alternative weighting, is 12.0 points — a very-high-HbA1c profile scores 12 points lower under cardiometabolic-heavy weighting than under the current weights, because that alternative lifts the Cardiometabolic domain from its current 34% to 50% and HbA1c is the metric carrying that profile down. The next largest is 10.0 points: a very-high-VO2max profile scores 10 points lower under equal-weighting, which cuts VO2 max from its current 0.42 share of Physical Capacity to a uniform 0.25 and the domain itself from 44% to 33%. Averaged across all 22 profiles, the mean absolute swing is 2.2–3.5 points for the three domain-level alternatives. Swapping the sleep/fiber emphasis inside Lifestyle is by far the smallest — at most 1 point, 0.2 on average — since it reorders two members inside the lightest domain rather than touching the domain weights themselves.
This is presented as a bound on the current weights' influence, not a defense that they are uniquely correct: a handful of reasonable people could disagree on the weighting and move scores by about 2.2–3.5 points on average, or up to 12.0 points in the worst case we tested. Worth holding next to a fact from the roadmap — one realistic lifestyle change is typically worth 1–4 points. For the profiles at the extremes, the choice of weighting can matter more than the change you make.