Baseline Survival DataCDC NCHS — United States Life Tables, 2022 (NVSR 74-02) Period life tables for US males and females. Provides the lx (survivors per 100,000) column by single year of age from 0 to 100+. This is the foundation of the survival curves.
Healthy Life ExpectancyJia & Lubetkin — Life expectancy and active life expectancy by disability status in older U.S. adults (PLOS ONE, 2020) Provides active life expectancy vs total life expectancy by age (65–95) and sex, using the National Health and Aging Trends Study. Key data: at age 65, US males have 12.1 active years out of 17.9 total (68% active). We use this to anchor the disability fraction curve for ages 65+.
GBD 2021 — Global Burden of Disease Study: HALE estimates (Lancet, 2024) Healthy life expectancy at birth for the US: 63.2 years for males, 65.7 for females (vs total LE of 74.3/80.0). Used to calibrate the disability onset curve for younger ages.
Mortality Risk Factors (Hazard Ratios)Global BMI Mortality Collaboration — BMI and all-cause mortality: 239 prospective studies, 4 continents (Lancet, 2016) The definitive BMI-mortality study. 10.6 million participants. U-shaped curve: BMI 22.5–25 = reference; 30–35 HR 1.44; 35–40 HR 1.88; 40+ HR 2.51; underweight (<18.5) HR 1.60.
Gellert et al — Smoking and all-cause mortality in older people: systematic review and meta-analysis (Arch Intern Med, 2012) Current smokers HR ~1.83–2.0 for all-cause mortality vs never-smokers. Former smokers (quit 5+ years) HR ~1.15.
Lee et al — Long-term leisure-time physical activity intensity and all-cause and cause-specific mortality (Circulation, 2022) 108,009 participants. Meeting guidelines (150–300 min/wk moderate): 20–25% lower mortality. 2–4× guidelines: 30–35% lower. Sedentary as reference.
Zhao et al — Association between daily alcohol intake and risk of all-cause mortality: a systematic review and meta-analyses (JAMA Network Open, 2023) Systematic review of 107 cohort studies. After correcting for abstainer bias, light drinking shows no mortality benefit (HR ~1.0). Heavy drinking (3+/day) HR 1.38.
Recovery After Lifestyle ChangeCho et al — Smoking Cessation and Short- and Longer-Term Mortality (NEJM Evidence, 2024) Excess mortality risk after quitting: ~50% remaining at 3 years, ~15% at 5–10 years, approaches never-smoker levels after 10+ years. We model this as exponential decay with half-life 3.5 years.
Xie et al — Association of Weight Loss Between Early Adulthood and Midlife With All-Cause Mortality Risk in the US (JAMA Network Open, 2020) Weight loss from obese to overweight associated with 54% mortality reduction. Benefits appear within 2–3 years. We model with half-life 2 years plus 15% residual risk from accumulated damage.
Mok et al — Physical activity trajectories and mortality: population based cohort study (BMJ, 2019) 14,599 participants in the EPIC-Norfolk cohort. Those who increased physical activity from inactive to moderately active had significantly lower mortality. Benefit appeared within the first few years of changing behavior. We model the transition as exponential decay with half-life 2 years.
Compression of Morbidity (Disability Shift Model)Chakravarty et al — Reduced disability and mortality among aging runners: a 21-year longitudinal study (Arch Intern Med, 2008) The landmark Stanford runners study. 538 runners vs matched controls, 21 years. Runners postponed disability by 12–16 years but death by only 3.3–7 years. This proves exercise compresses morbidity — the disabled period shrinks. We use a disability shift multiplier of 2.3× for exercise.
Fries et al — Compression of Morbidity 1980–2011: A Focused Review of Paradigms and Progress (Journal of Aging Research, 2011) Reviews 30 years of evidence. Low-risk cohorts (no smoking, no obesity, exercising) postponed disability by 8.3 years vs death by 3.6 years. Ratio ~2.3:1. "Postponement of disability is several-fold the postponement of mortality."
Klijs et al — Obesity, smoking, alcohol and years lived with disability: a Sullivan life table approach (BMC Public Health, 2011) Dutch cohort, n=6,446. Critical finding: each factor affects disability differently than mortality. Obese: 5.9 yrs disabled vs 3.2 for normal weight (LE diff only 1.1 yrs) — disability shift 3.0×. Smokers: 3.8 yrs disabled regardless of status — disability shift ~1.0×. This is why we use per-factor disability multipliers.
Reuser et al — "Smoking kills, obesity disables" — a multistate approach (Obesity, 2009) US Health and Retirement Survey analysis confirming that smoking primarily affects mortality while obesity primarily affects disability, with fundamentally different pathways through the healthy→disabled→dead transitions.
Strength TrainingShailendra et al — Resistance Training and Mortality Risk: A Systematic Review and Meta-Analysis (Am J Prev Med, 2022) Meta-analysis of 10 studies. Any resistance training reduced all-cause mortality by 15% (RR 0.85, 95% CI 0.77–0.93) independent of aerobic exercise. This is the basis for our mortality HR of 0.85 for regular strength training.
Shailendra et al — Weight training and risk of all-cause, cardiovascular disease and cancer mortality among older adults (Int J Epidemiol, 2024) NIH-AARP Diet and Health Study. Any weight training: HR 0.94 for all-cause mortality, 0.92 for CVD, 0.95 for cancer — after adjusting for aerobic exercise. Confirms the independent, additive effect of strength training.
Law, Clark & Clark — Resistance Exercise to Prevent and Manage Sarcopenia and Dynapenia (Ann Rev Gerontol Geriatr, 2016) Sarcopenia (age-related muscle loss) is the #1 cause of functional dependence in older adults — prevalence 10% at age 60, over 50% at 80+. Resistance training is the most effective intervention. This is why we use a disability shift multiplier of 3.0× for strength training — its effect on staying independent far exceeds its effect on mortality.
Fragala et al — Resistance Training for Older Adults: Position Statement From the NSCA (J Strength Cond Res, 2019) Comprehensive position statement. Progressive resistance training increases muscle strength, muscle size, and functional capacity. Directly prevents the loss of independence that defines the transition from "healthy years" to "impaired years."
Sleep DurationCappuccio et al — Sleep Duration and All-Cause Mortality: A Systematic Review and Meta-Analysis (Sleep, 2010) Meta-analysis of 16 studies covering 1.38 million participants. Short sleep (<6 hrs) HR 1.12 for all-cause mortality. Long sleep (>8 hrs) HR 1.30. The U-shaped relationship is consistent across populations. We use <5 hrs HR 1.21, 5–6 hrs HR 1.14, 7–8 hrs as reference, 9+ hrs HR 1.34.
Itani et al — Short sleep duration and health outcomes: a systematic review, meta-analysis, and meta-regression (Sleep Medicine, 2017) Confirmed short sleep duration is significantly associated with mortality, diabetes, hypertension, cardiovascular disease, and obesity. Short sleep is both a mortality risk and a disability accelerator through its effects on cognitive function and metabolic health.
Social ConnectionHolt-Lunstad et al — Loneliness and Social Isolation as Risk Factors for Mortality: A Meta-Analytic Review (Perspectives on Psychological Science, 2015) Landmark meta-analysis of 70 studies covering 3.4 million participants. Social isolation HR 1.29, loneliness HR 1.26 for all-cause mortality. The effect size is comparable to smoking 15 cigarettes per day. This is the basis for our isolation HR of 1.29.
Livingston et al — Dementia prevention, intervention, and care (Lancet, 2020) Social isolation is identified as one of 12 modifiable risk factors for dementia, increasing risk by approximately 60%. This is why we apply a disability shift multiplier of 2.0× for social isolation — its impact on cognitive decline and functional independence is substantially larger than its mortality effect. Maintaining social connections is one of the strongest interventions for preserving independence in later life.
International Life TablesWHO Global Health Observatory — Life tables by country, 2021 Abridged life tables (nqx — probability of dying in each age interval) for 183 WHO Member States by sex, computed using standardized WHO methodology. We interpolate to single-year-of-age lx values using log-linear interpolation within each age interval. These are period life tables reflecting 2021 mortality conditions.
WHO Global Health Observatory — Healthy Life Expectancy (HALE) at birth, 2021 HALE estimates for all countries, by sex. We compute the HALE/LE ratio for each country and use it to scale our age-specific healthy fraction curves. Countries with higher HALE/LE ratios (e.g. Japan, Singapore) have their disability onset shifted later; countries with lower ratios have it shifted earlier. The US baseline healthy fraction curve (derived from NHIS and PLOS data above) serves as the reference.
Global BMI Mortality Collaboration — BMI and all-cause mortality across 4 continents (Lancet, 2016) Confirms that BMI-mortality hazard ratios are "broadly similar across geographical regions" (Europe, North America, East Asia, Australia/NZ). This validates using the same lifestyle hazard ratios across all countries — we only need to swap the baseline survival curves, not the risk factor model.
Good Years Left