← Back

Methodology

How Lumenai computes the numbers you see. Every score is grounded in peer-reviewed evidence, with the source cited.

Biological age

Lumenai Biological Age

Your biological age is one number that integrates multiple peer-reviewed dimensions of aging into a single, transparent estimate. PhenoAge (Levine et al., 2018) is the anchor — a validated lab-based estimator of biological age. We then apply bounded lifestyle adjustments grounded in independent mortality literature for cardiorespiratory fitness, resting heart rate, body composition, HRV trend, and sleep.

Each adjustment has a defined range, a source citation, and a visible contribution on your dashboard. The total adjustment is bounded to ±5 years so a single dimension can't override the others. PhenoAge itself is unconstrained — it speaks the truth of your blood work.

When wearable or body-composition data isn't available, your score degrades gracefully to the pure PhenoAge anchor with a "limited data" indicator. Adding more data sources increases the precision and unlocks more accurate adjustments.

The lifestyle adjustments

AdjustmentRangeSource
Cardiorespiratory fitness
VO₂max percentile vs Cooper norms
−2.5 to +3.0Mandsager 2018, JAMA Netw Open
Resting heart rate
90-day average bpm
−0.5 to +1.0Fox 2008, NEJM (BEAUTIFUL)
Body composition
Body fat % vs age/sex norms
−0.5 to +1.0Tomiyama 2016, Int J Obes
HRV trend
90-day vs your own baseline
−0.3 to +0.5Shaffer & Ginsberg 2017
Sleep duration + quality
30-day average + sleep score
−0.2 to +0.6Cappuccio 2010, Sleep

The PhenoAge anchor

PhenoAge (Levine et al., 2018) is validated against all-cause mortality in the NHANES dataset and consistently outperforms chronological age as a predictor of healthspan.

The model uses nine standard biomarkers from any comprehensive blood panel, plus your chronological age:

  • · Albumin — liver synthesis and overall nutritional status
  • · Creatinine — kidney function and muscle metabolism
  • · Glucose (fasting) — metabolic baseline
  • · C-reactive protein — systemic inflammation
  • · Lymphocyte percentage — immune system composition
  • · Mean corpuscular volume (MCV) — red cell size
  • · Red cell distribution width (RDW) — independent mortality predictor
  • · Alkaline phosphatase — liver, bone, intestinal markers
  • · White blood cell count — inflammation and immune state

Each marker contributes proportionally to a linear combination that maps to a 10-year mortality probability via a Gompertz proportional hazards model, which is then transformed back to an age estimate. Lower than your chronological age means your biomarkers look younger than your years; higher means older.

Source. Levine ME, Lu AT, Quach A, et al. An epigenetic biomarker of aging for lifespan and healthspan. Proc Natl Acad Sci USA. 2018;115(34):8329–8334.View paper
Health score

Lumenai Health Score

The Health Score is a transparent composite from 0 to 100, computed as a weighted average across marker categories. Each category produces a sub-score from 0 to 100 based on how many of your tracked markers in that category sit in the optimal, watch, or flag bands.

Per-marker contribution

Each marker in a category contributes to that category's sub-score:

  • · Optimal — 100 points
  • · Watch — 60 points
  • · Flag — 20 points
  • · Missing markers don't penalize the score; they simply aren't counted

Category weights

Categories are weighted by their contribution to long-term mortality and disability-adjusted life years, informed by the Global Burden of Disease Study 2019 (Institute for Health Metrics and Evaluation). Weights are renormalized to categories you have data for, so you're never penalized for not running a marker.

CategoryWeightWhy
Cardiovascular22%Leading cause of mortality globally (GBD 2019). Includes lipid profile, ApoB, Lp(a), homocysteine, inflammation.
Metabolic18%Drives downstream cardiovascular, cognitive, and cancer risk. Includes glucose, HbA1c, insulin, HOMA-IR.
fitness13%Cardiorespiratory fitness is the strongest single predictor of all-cause mortality (Mandsager et al., JAMA Netw Open 2018).
Hormones10%Cross-system effects on metabolism, mood, recovery, body composition.
Inflammation8%hsCRP independently predicts cardiovascular events (Ridker et al., NEJM 2002).
Kidney6%CKD prevalence ~10% globally; irreversible if missed early. KDIGO 2024.
Vitamins6%Broadly modifiable, broad health effects. Vitamin D, B12, folate.
Liver5%NAFLD prevalence ~25% globally; rising with metabolic syndrome.
body_composition3%Body composition independent of BMI (Tomiyama et al., Int J Obes 2016).
Thyroid3%Subclinical thyroid dysfunction is common and modifiable.
Bone3%Osteoporosis screening, especially post-menopause.
Cancer screen2%Screening signal only — informative but not diagnostic.
Minerals1%Trace elements (zinc, selenium, magnesium) and iron status.

Note: blood count markers (CBC) are not weighted in the Health Score because they already feed into PhenoAge directly. Weighting them here would double-count.

A note on weights. The weights above represent Lumenai's clinical judgment informed by population mortality data. They are deliberately public so you can see exactly how your score is built. We may refine them over time as the evidence base matures and as the medical advisory board grows. Any changes will be announced.
Marker status

Optimal vs watch vs flag

For every marker in your record, Lumenai classifies your value as optimal, watch, or flag against ranges sourced from peer-reviewed literature — not the broad reference ranges most labs report. Reference ranges are designed to capture 95% of the population and miss most subclinical risk. Optimal ranges are designed to capture healthy.

Where age or sex meaningfully changes the expected range — testosterone, eGFR, ferritin, AMH, PSA — the optimal band is stratified accordingly. A 32-year-old man and a 58-year-old man have different optimal testosterone ranges, and Lumenai reflects that.

Where the evidence is mature, we cite a specific source on the marker detail page (Endocrine Society 2018, AHA/ACC 2018, ESC/EAS 2019, ATA 2014, KDIGO 2024). Where peer-reviewed consensus on "optimal" doesn't exist, we use the lab reference range and flag this in the UI rather than invent a number.

Our promise

What we don't do

Lumenai doesn't diagnose. It doesn't prescribe. It synthesises your data, flags what's worth attention, and tells you what the evidence says about each marker. Final clinical decisions rest with you and your physician.

We also don't hide our methodology behind an opaque score. If you ever want to see exactly why a number is what it is, ask the chat — "how is my PhenoAge computed?" or "why is my cardiovascular sub-score 72?" — and you'll get a marker-by-marker explanation, not a black box.