Outliyr · Methodology · v2.0.0
How the biological age estimate works.
Every answer is converted into age-equivalent years using published mortality research, so across the adult population the results average out at calendar age. With a complete blood panel, a lab-validated PhenoAge is blended in. Every number and source is listed below.
← Back to the calculatorHow a result is calculated
- Each answer has a published risk. For every question we use the adjusted all-cause mortality hazard ratio (HR) from large cohort studies or meta-analyses, cited in the tables below.
- Risk becomes years. Adult mortality risk rises about 9.2% per year of age (it doubles roughly every 7.9 years). So a hazard ratio converts to age-equivalent years as ln(HR) ÷ 0.0880.
- Average answers score zero. Each question is centred on how common each answer is in the adult population, so the average answer adds 0 years and the population as a whole averages out at calendar age. Risk is lopsided (a minority smoke, are sedentary, or are highly stressed), so slightly more than half of people score a little younger.
- Overlap is removed. Many habits travel together (fit people tend to sleep better and smoke less), so single-factor effects overstate the combined effect. Each contribution is scaled down before adding them up: by 0.6 for lifestyle, fitness and symptom questions, and by 0.8 for smoking and lab values, which overlap less with everything else. No single answer can move the result more than 6 years.
- Your 8 systems show where your years come from. Each answer belongs to one system. A system’s age shows how unusual that system is, on the same scale as the overall result.
- Labs, when you have them. With all nine PhenoAge markers (albumin, creatinine, fasting glucose, hs-CRP, lymphocyte %, MCV, RDW, alkaline phosphatase, white cell count) the result blends in PhenoAge (Levine et al., 2018) at 60%, and those markers are not counted a second time.
Validation status and intended use
The individual effect sizes come from peer-reviewed mortality studies, and PhenoAge is a validated blood-based age. The combined questionnaire model has not itself been validated in a prospective cohort. Self-reported answers are imprecise, and some questions (marked “prior” below) have little direct mortality evidence, so they are given small, conservative effects.
Use the result to see which areas are most likely costing or saving you years, choose one change to test, and compare future results under the same model version. It does not diagnose disease or replace laboratory interpretation by a qualified clinician.
Cardiovascular
Model version: 2.0.0
| Input | Age-equivalent years (after overlap adjustment) | Evidence |
|---|---|---|
| Stair climb / VO2max Questionnaire · vo2max | not winded: -0.5 slightly winded: +0.5 very winded: +3.1 can't finish: +4.1 Exact value, if you know it: 15 mL/kg/min: +4.9 25 mL/kg/min: +2.2 35 mL/kg/min: -0.5 45 mL/kg/min: -3.2 55 mL/kg/min: -5.9 | Moderate · slightly winded 1.16, very winded 1.70, cannot finish 1.96; exact VO2max 0.87 per MET Dyspnea: ARIC (mMRC 1-2 HR 1.16, mMRC 3-4 1.96); fitness: Kodama 2009 JAMA meta (low vs high CRF 1.70; 0.87 per MET) PMID:27780208, PMID:19454641 |
| Exercise per week Questionnaire · exercise_minutes_week | rarely or never: +1.7 1-2 days/week: +0.2 3-4 days/week: -0.8 5+ days/week: -1.5 Exact value, if you know it: 0 min/week: +1.7 75 min/week: +0.2 150 min/week: -0.8 300 min/week: -1.4 600 min/week: -1.6 | Strong · vs none: below guideline 0.80, 1-2x guideline 0.69, 2-5x 0.61-0.63; no harm at 10x Arem 2015 JAMA Intern Med, pooled 661,137 adults, leisure-time activity dose-response PMID:25844730 |
| Smoking Questionnaire · smoker | never: -1.7 former: +0.3 current: +6.0 | Strong · current vs never: women 3.0, men 2.8 (10+ years of life lost); former: most excess removed by quitting (prior +2.5 y) Jha 2013 NEJM (NHIS, 202,248 adults) PMID:23343063 |
| Resting heart rate Wearable · resting_hr_avg | under 60: -1.4 60-80: +0.3 over 80: +1.9 Exact value, if you know it: 45 bpm: -2.4 55 bpm: -1.4 70 bpm: +0.3 80 bpm: +1.3 90 bpm: +2.4 100 bpm: +3.5 | Strong · 1.17 per 10 bpm Aune 2017 Nutr Metab Cardiovasc Dis meta, 87 studies PMID:28552551 |
| Blood pressure Questionnaire · bp_self_report | normal: -0.4 low: -0.4 high / on meds: +1.1 don't know: +0.2 | Strong · untreated hypertension 1.40, treated uncontrolled 1.62, controlled = normal; 'high/on meds' blended +2.5 y; 'don't know' prior +1 y NHANES III linked mortality, 13,947 adults, 19 y PMID:29925884 |
| ApoB Bloodwork · apob | 60 mg/dL: -0.3 100 mg/dL: -0.3 130 mg/dL: +0.4 160 mg/dL: +1.0 | Weak · above ~100 mg/dL rises ~1.08 per 30 mg/dL NHANES ApoB cohort: CV mortality 1.13 per SD; all-cause U-shaped (no reward for very low) PMID:37611866 |
| Triglycerides Bloodwork · triglycerides | 90 mg/dL: -0.4 149 mg/dL: -0.4 175 mg/dL: +0.4 250 mg/dL: +1.3 400 mg/dL: +2.3 | Moderate · 150-199: 1.09, >=200: 1.20 vs 90-149 mg/dL; +12% per mmol/L Liu 2013 Lipids Health Dis PMID:24164719 |
| Heart rate variability (RMSSD) Wearable · hrv_avg | 15 ms RMSSD: +2.3 25 ms RMSSD: +0.8 35 ms RMSSD: -0.8 60 ms RMSSD: -0.8 | Moderate · lowest RMSSD quartile vs others 1.56 Jarczok 2022 Neurosci Biobehav Rev meta, 38,008 participants PMID:36243195 |
| Daily steps Wearable · steps_avg | 3553 steps/day: +3.2 5801 steps/day: -0.3 7842 steps/day: -0.9 10901 steps/day: -2.0 | Strong · vs ~3,500/day: ~5,800 0.60, ~7,800 0.55, ~10,900 0.47 Paluch 2022 Lancet Public Health, 15 cohorts PMID:35247352 |
Metabolic
Model version: 2.0.0
| Input | Age-equivalent years (after overlap adjustment) | Evidence |
|---|---|---|
| Waist / pants fit Questionnaire · waist_cm | loose: -0.2 fits well: -0.2 tight / going up a size: +0.4 Exact value, if you know it: 80 cm: -0.2 90 cm: +0.4 100 cm: +0.9 110 cm: +1.4 120 cm: +1.9 | Strong · per 5 cm: men 1.07, women 1.09; 'tight/going up a size' ~ +5 cm; looser is not rewarded (weight loss can signal illness) Cerhan 2014 Mayo Clin Proc, 650,386 adults, BMI-adjusted PMID:24582192 |
| Going 4 hours without food Questionnaire · fasting_tolerance_q | comfortable: -0.1 mild: -0.1 jittery: +0.2 can't skip meals: +0.5 | Prior · no general-population mortality data; kept at or under ~1 year No solid mortality evidence; conservative prior |
| Crash after a carb-heavy meal Questionnaire · post_meal_crash_q | never: -0.1 sometimes: -0.1 often: +0.2 always: +0.5 | Weak · prior: often +0.5 y, always +1 y Bergman 2016 (elevated 1-h glucose ~28% higher mortality; symptom not validated against glucose) PMID:26996391 |
| Fasting glucose Bloodwork · fasting_glucose | 70 mg/dL: -0.9 99 mg/dL: -0.9 110 mg/dL: +0.4 126 mg/dL: +4.5 180 mg/dL: +6.0 | Strong · diabetes vs none 1.80; fasting glucose above 5.6 mmol/L (100 mg/dL) associated with death Emerging Risk Factors Collaboration 2011 NEJM PMID:21366474 |
| HbA1c Bloodwork · hba1c | 5 %: -0.8 5.6 %: -0.8 6 %: +0.1 6.5 %: +2.3 8 %: +4.5 | Moderate · J-shaped; rises from 5.7%; 1.4-1.8 in the diabetic range; no reward for very low Selvin 2010 NEJM (ARIC); ERFC 2011 for diabetes PMID:20200384 |
| HOMA-IR (insulin resistance) Bloodwork · homa_ir | 1: -0.9 2: -0.0 3: +0.8 5: +1.8 | Moderate · highest vs lowest HOMA-IR 1.34 Biosci Rep 2017 meta, 26,976 adults PMID:28811358 |
| Albumin Bloodwork · albumin | 3.5 g/dL: +2.5 3.8 g/dL: +1.1 4.1 g/dL: -0.2 4.6 g/dL: -0.2 | Moderate · 1.16 per SD lower; used inside PhenoAge when the full panel is present Moli-sani cohort, adults 35+ PMID:39010980 |
Inflammatory
Model version: 2.0.0
| Input | Age-equivalent years (after overlap adjustment) | Evidence |
|---|---|---|
| hs-CRP Bloodwork · hs_crp | 0.5 mg/L: -4.1 1 mg/L: -1.3 3 mg/L: +2.7 10 mg/L: +6.0 | Strong · non-vascular death 1.54 per ~3-fold higher CRP (log-linear) Emerging Risk Factors Collaboration 2010 Lancet PMID:20031199 |
| Homocysteine Bloodwork · homocysteine | 7 umol/L: -2.1 10 umol/L: -0.5 15 umol/L: +2.1 20 umol/L: +4.8 | Strong · 1.34 per 5 umol/L Fan 2017 Sci Rep meta PMID:28684797 |
| Ferritin Bloodwork · ferritin | 30 ug/L: -0.4 200 ug/L: -0.4 300 ug/L: +0.5 600 ug/L: +3.3 | Moderate · >=200 vs <200: 1.1; >=600: 1.5 (low end not scored) Ferritin cohort PMID:25156997 |
| RDW Bloodwork · rdw | 12.5 %: -1.4 13.5 %: +0.5 14.5 %: +2.3 16 %: +4.9 | Strong · 1.23 per 1 SD (~1%); used inside PhenoAge when the full panel is present Perlstein 2009 Arch Intern Med, NHANES III PMID:19307522 |
| Joint and body pain Questionnaire · joint_pain_q | 0: -0.5 1: -0.5 2: -0.5 3: -0.4 4: -0.3 5: -0.2 6: +0.6 7: +1.4 8: +1.8 9: +2.2 10: +2.6 | Moderate · mild pain ~no effect once adjusted; pain limiting activity 1.33; severe widespread 1.59 HRS 1998-2018 (pain interference 1.33 fully adjusted); Macfarlane 2017 UK Biobank + meta (chronic widespread pain 1.59) PMID:41520747, PMID:28733474 |
| Seasonal allergies Questionnaire · allergy_burden_q | none: -0.0 mild: -0.0 moderate: -0.0 severe: +0.3 | Weak · allergies are neutral-to-protective, so not penalised; severe (possible asthma overlap) prior +0.5 y Konishi 2019 J Epidemiol (pollinosis HR 0.79); Crans Yoon 2016 (allergic rhinitis 0.42) PMID:29937471, PMID:27742084 |
| Recovery after 2 drinks Questionnaire · alcohol_recovery_q | fine next morning: -0.0 mild hangover: -0.0 bad hangover: +0.3 don't drink: -0.0 | Prior · bad hangover prior +0.5 y; no credit for drinking Zhao 2023 JAMA Netw Open (107 cohorts): risk rises from ~45 g/day; hangover severity is only a loose proxy PMID:37000449 |
Immune
Model version: 2.0.0
| Input | Age-equivalent years (after overlap adjustment) | Evidence |
|---|---|---|
| Colds or flu per year Questionnaire · cold_freq_q | 0 a year: -0.1 1-2 a year: -0.1 3+ a year: +0.2 | Prior · adults average 2-3 colds a year; 3+ prior +0.5 y No solid mortality evidence; conservative prior |
| Recovery time when sick Questionnaire · cold_recovery_q | 1-3 days: -0.1 about a week: -0.1 2+ weeks: +0.5 | Prior · 2+ weeks to recover prior +1 y No solid mortality evidence; conservative prior |
| Cold sores / shingles Questionnaire · viral_reactivation_q | never: -0.0 rare: -0.0 often: +0.3 | Prior · frequent cold sores/shingles prior +0.5 y No solid mortality evidence; conservative prior |
| Neutrophil-to-lymphocyte ratio Bloodwork · neutrophil_lymphocyte_ratio | 1.2: -1.8 1.7: -0.9 2.1: +0.3 2.5: +1.6 3.5: +3.6 | Moderate · top vs bottom fifth (NLR > 2.4) 1.64 Fest 2019 Rotterdam Study PMID:30569368 |
| Vitamin D (25-OH) Bloodwork · vitamin_d | 11 ng/mL: +2.0 16 ng/mL: +1.0 21 ng/mL: +0.2 31 ng/mL: -0.5 35 ng/mL: -0.5 | Moderate · from ~11 ng/mL: +5 ng/mL 0.86, +10 0.77, +20 0.69, plateau above ~30 Zittermann 2012 Am J Clin Nutr meta (observational; supplementation trials null) PMID:22170374 |
Musculoskeletal
Model version: 2.0.0
| Input | Age-equivalent years (after overlap adjustment) | Evidence |
|---|---|---|
| Standing up from a chair Questionnaire · chair_stand_reps | easy: -0.9 some effort: +0.3 hard: +1.5 need hands: +2.7 Exact value, if you know it: 5 reps/30s: +2.7 10 reps/30s: +1.5 14 reps/30s: +0.3 18 reps/30s: -0.9 25 reps/30s: -0.9 | Moderate · capped at Cooper-scale effects: some effort +2, hard +4, needs hands +6 years before overlap shrinkage Cooper 2010 BMJ meta (physical capability incl. chair rise; grip weakest vs strongest 1.67); Brito 2014 sit-rise (direction, HRs inflated) PMID:20829298, PMID:23242910 |
| Push-ups Questionnaire · pushup_reps | 20+ easily: -0.1 10-20: -0.1 under 10: +0.2 can't do any: +0.5 Exact value, if you know it: 0 reps: +0.5 10 reps: +0.2 20 reps: -0.1 | Weak · prior: few +0.5 y, none +1 y Yang 2019 JAMA Netw Open (CVD events, 1,104 firefighters; not mortality) PMID:30768197 |
| Walking pace vs peers Questionnaire · stair_pace_q | faster: -0.3 same: +0.0 slower: +2.1 | Strong · brisk vs average ~0.95; slow vs average ~1.35 (conservative between Stamatakis 1.25 and Yates 1.3-2.2) Stamatakis 2018 BJSM (50,225 walkers); Yates 2017 Eur Heart J (UK Biobank) PMID:29858463, PMID:29020281 |
Neuro-cognitive
Model version: 2.0.0
| Input | Age-equivalent years (after overlap adjustment) | Evidence |
|---|---|---|
| Mental sharpness Questionnaire · cognition_q | 0: +1.0 1: +1.0 2: +1.0 3: +0.4 4: +0.4 5: -0.2 6: -0.2 7: -0.2 8: -0.2 9: -0.2 10: -0.2 | Weak · average or better: no effect (no evidence that sharper-than-average buys years); 3-4 +1 y, 0-2 +2 y HUNT: subjective short-term memory impairment 1.51, strongest at 70-79 (scaled down for younger adults) PMID:31825300 |
| Brain fog Questionnaire · brain_fog_q | never: -0.1 rarely: -0.1 often: +0.2 daily: +0.5 | Prior · overlaps energy and sharpness; often +0.5 y, daily +1 y No solid mortality evidence; conservative prior |
| Recalling names and words Questionnaire · recall_q | 0: +2.5 1: +2.0 2: +1.6 3: +1.1 4: +0.4 5: -0.4 6: -0.4 7: -0.4 8: -0.4 9: -0.4 10: -0.4 | Moderate · 0 words 1.52, 1 word 1.24 vs 3 words (elderly only) 3-word delayed recall, 3,778 non-demented older adults, 10 y PMID:20652904 |
Hormonal / Stress
Model version: 2.0.0
| Input | Age-equivalent years (after overlap adjustment) | Evidence |
|---|---|---|
| Stress this month Questionnaire · wired_tired_q | calm: -0.8 some, manageable: +0.4 high, often overwhelmed: +1.6 constantly overwhelmed: +3.7 | Strong · some 1.20, high 1.43, constant 1.94 vs none Russ 2012 BMJ, 68,222 adults (GHQ-12 psychological distress) PMID:22849956 |
| Energy on a typical day Questionnaire · fatigue_q | 0: +1.7 1: +1.6 2: +1.4 3: +1.0 4: +0.5 5: +0.1 6: -0.2 7: -0.5 8: -0.9 9: -0.9 10: -0.9 | Moderate · most-fatigued quarter 1.40 vs most-vital quarter EPIC-Norfolk vitality, 18,101 adults, 16.6 y (BMC Med 2016) PMID:27543008 |
| Daily energy pattern Questionnaire · energy_curve_q | high in the morning: -0.3 steady: -0.3 peaks in the evening: +0.4 always exhausted: +1.3 | Moderate · evening type 1.10; always exhausted 1.26 Knutson & von Schantz 2018 Chronobiol Int, UK Biobank 433,268 (definite evening type 1.10); exhaustion mapped to EPIC-Norfolk fatigue (1.26, fully adjusted) PMID:29642757, PMID:27543008 |
| Afternoon energy Questionnaire · afternoon_dip_q | strong: -0.1 brief dip: -0.1 major crash: +0.6 | Weak · brief dip no effect; major crash ~1.10 Wang 2024 PLoS ONE meta (naps under 1 h 1.00; 1 h+ 1.22) PMID:39413101 |
Sleep / Circadian
Model version: 2.0.0
| Input | Age-equivalent years (after overlap adjustment) | Evidence |
|---|---|---|
| Sleep quality Questionnaire · sleep_quality_q | 0: +1.1 1: +1.0 2: +0.9 3: +0.8 4: +0.4 5: +0.1 6: -0.1 7: -0.3 8: -0.3 9: -0.3 10: -0.3 | Moderate · non-restorative sleep 1.23, difficulty falling asleep 1.13 Ge 2019 Sleep Med Rev meta, 29 cohorts, 1.6M adults PMID:31630016 |
| Sleep duration Questionnaire · sleep_duration_hours_q | under 5 hours: +0.2 5-6 hours: +0.1 7-9 hours: -0.2 over 9 hours: +1.6 Exact value, if you know it: 4 hours: +0.1 5 hours: +0.2 6 hours: +0.1 7 hours: -0.2 8 hours: -0.0 9 hours: +0.6 10 hours: +1.3 11 hours: +2.0 | Strong · vs 7 h: 5 h 1.06, 6 h 1.04, 9 h 1.13, 10 h 1.25 (long sleep partly reverse causation) Liu 2017 Sleep Med Rev meta-regression (40 cohorts, 2.2M); Cappuccio 2010 Sleep PMID:27067616, PMID:20469800 |
| Measured sleep (wearable) Wearable · sleep_total_avg | 4 hours: +0.1 5 hours: +0.2 6 hours: +0.0 7 hours: -0.2 8 hours: -0.0 9 hours: +0.6 10 hours: +1.3 11 hours: +2.0 | Strong · same curve as sleep duration Liu 2017 Sleep Med Rev (wearable-measured hours replace the self-reported answer) PMID:27067616 |
| Bedtime consistency Questionnaire · sleep_consistency_q | always (±30 min): -0.5 most nights: -0.2 rarely: +0.4 never: +1.3 | Strong · mapped conservatively: very consistent -0.5 y, irregular +1 y, very irregular +2.5 y Windred 2024 Sleep (fully adjusted top vs bottom sleep-regularity 0.70); Cribb 2023 eLife PMID:37738616, PMID:37995126 |
