Methodology & Validation
A life expectancy estimate is only as good as its agreement with reality. We benchmark the mortality engine against published clinical literature — peer-reviewed survival studies, government databases, and professional-society reports. Of 116 benchmark cases, 109 (94%) land inside the published target range, and all 116 fall within 1.5 years of it. Every case below is live engine output, shown against its source range and its Monte Carlo confidence interval.
How we validate
Each benchmark case runs a real applicant profile through the engine's survival model and extracts the remaining life expectancy. That estimate is compared against a target range derived from published clinical literature — peer-reviewed journal articles (with DOIs), government databases (SSA, SEER, USRDS), and professional-society reports (SOA, Alzheimer's Association, ISHLT). We also report the Monte Carlo 90% confidence interval (5th–95th percentile of simulated outcomes) so the uncertainty around each estimate is visible, not hidden.
Verdicts are deliberately strict: within range means the engine estimate falls inside the published target band; close means within 1.5 years of a boundary; anything further is a miss. The baselines are validated first (a healthy 65-year-old must match SSA life tables before any condition is layered on), so impairment results build on a verified foundation.
How to read each bar
What the benchmarks show
Every row is generated by the production engine — nothing here is hand-written. The per-carrier underwriting estimator and the settlement valuation engine build on this same validated mortality core.
1. Healthy Baselines
Validates VBT 2015 Ultimate tables with MP-2021 mortality improvement projection against SSA period life tables.
Scale: 0–35 years remaining
SSA Period Life Table 2021 (male age 65: 18.0yr); VBT with improvement projects higher [source]
SSA Period Life Table 2021 (male age 70: 14.7yr); SOA 2015 VBT Ultimate Table 3269 [source]
SSA Period Life Table 2021 (male age 75: 11.6yr); VBT with improvement projects higher [source]
SSA Period Life Table 2021 (male age 80: 8.7yr) [source]
SSA Period Life Table 2021 (male age 85: 5.9yr) [source]
SSA Period Life Table 2021 (female age 70: 16.8yr) [source]
SSA Period Life Table 2021 (female age 75: 13.5yr) [source]
SSA Period Life Table 2021 (female age 80: 10.1yr) [source]
2. Cardiovascular Conditions
CHF, CAD, and cardiomyopathy validated against JACC meta-analyses, Medicare claims data, and Framingham Heart Study.
Scale: 0–25 years remaining
Taylor CJ et al., BMJ Open 2019: CHF 5-year mortality 50-75% depending on EF. [source]
Taylor CJ et al., BMJ Open 2019. Medicare data: median survival ~5yr at 75. [source]
Medicare CHF cohort: median survival ~5yr at age 75. [source]
CHF at 80+: median survival 2.5-4yr. Jhund PS et al., JAMA 2009. [source]
NYHA IV 1-year mortality 50-75%. SOLVD, Val-HeFT trials. [source]
NYHA IV at 80: extremely poor prognosis. MAGGIC meta-analysis. [source]
HFrEF with LVEF <25%: 5-year mortality ~60-70%. MAGGIC meta-analysis. [source]
DCM 5-year survival ~50%. Felker GM et al., NEJM 2000. [source]
DCM at older ages: worse prognosis. Felker GM et al., NEJM 2000. [source]
Untreated severe AS: median survival 2-3yr. Ross J, Braunwald E, Circulation 1968. [source]
Severe AS at 80: median 1-2yr untreated. PARTNER trial data. [source]
3. Respiratory Conditions
COPD validated against GOLD staging data and TORCH trial results. IPF against Ley B et al. Severe asthma against GINA 2023 guidelines.
Scale: 0–35 years remaining
GOLD 3-4: 5-year mortality 40-70%. TORCH trial. Shavelle RM, LE estimates 2009. [source]
TORCH trial (Calverley PM et al., NEJM 2007). [source]
COPD GOLD 3-4 at 80+: very poor prognosis. BODE index studies. [source]
GOLD 4 (FEV1 <30%): median survival ~5yr. BODE index studies. [source]
GOLD 4 at 80: very limited survival. [source]
IPF median survival 3-5yr from diagnosis. Ley B et al., AJRCCM 2011. [source]
IPF at 75: age at diagnosis worsens prognosis. [source]
GINA 2023: well-controlled severe asthma ~1.3x mortality. Engelkes C et al., Eur Respir J 2015. [source]
Asthma-COPD overlap: worse outcomes than either alone. GINA/GOLD 2023. Alshabanat A et al., PLoS One 2015. [source]
4. Metabolic Conditions
Diabetes and CKD validated against Emerging Risk Factors Collaboration, USRDS data, and Go et al. (NEJM 2004).
Scale: 0–30 years remaining
ERFC: DM2 reduces LE by 3-6yr. Seshasai S et al., NEJM 2011. [source]
ERFC: DM2 reduces LE by 3-6yr at 70. [source]
DM2 LE impact attenuates at advanced age. ERFC. [source]
HbA1c >9%: 2-3x mortality increase. UKPDS, ACCORD trials. [source]
CKD stage 4: HR ~3.2 vs normal. Go AS et al., NEJM 2004. [source]
CKD stage 4 at 75. Go AS et al., NEJM 2004. [source]
USRDS 2023: 5-year survival ~35% for dialysis age 65-74. [source]
USRDS 2023: worse outcomes at 75+. [source]
USRDS 2023: dialysis at 80+ median survival ~2yr. [source]
5. Oncology Conditions
Cancer validated against SEER 5-year relative survival rates (2013-2019 cohort). Multi-age testing for key cancer types.
Scale: 0–30 years remaining
SEER: Stage III NSCLC 5-year survival 10-36%. [source]
SEER: Stage III NSCLC. Worse outcomes at older ages. [source]
SEER: Stage IV NSCLC 5-year survival 0-10%. Median ~4-6 months. [source]
SEER: Stage IV pancreatic median survival 3-6 months. [source]
SEER: Stage IV pancreatic at 75. [source]
SEER: Stage III colorectal 5-year survival 53-65%. [source]
SEER: Stage I-II prostate 5-year survival ~99%. [source]
SEER: Stage IV prostate 5-year survival ~31%. [source]
SEER: Differentiated thyroid 5-year survival >98%. [source]
Glioblastoma median survival 12-18mo. Stupp R et al., NEJM 2005. [source]
GBM at 75: median ~6-9mo. Age is strongest prognostic factor. [source]
ETERNITY (EORTC 1419): median OS 8.92yr for recurrent IDH-wt GBM 5+ yr survivors. [source]
SEER: 5-yr conditional 5-yr survival >90% for pancreatic cancer survivors. [source]
SEER: conditional 5-yr survival 90.2% at 4yr. Hazard stabilizes after 2yr. [source]
6. Neurological Conditions
Dementia and ALS validated against Alzheimer's Association reports, ALS registries, and population cohort studies.
Scale: 0–45 years remaining
Alzheimer's Association 2024: median survival 4-8yr from diagnosis. [source]
Todd S et al., Int Psychogeriatr 2013. [source]
Later onset dementia: shorter survival. Xie J et al., BMJ 2008. [source]
Severe dementia: median survival 1-3yr. Mitchell SL et al., NEJM 2009. [source]
ALS median survival 2-5yr. Chio A et al., Neurology 2009. [source]
Older onset ALS: shorter survival. Calvo A et al., J Neurol 2017. [source]
PD SMR ~1.5-2.0. LE reduction ~5yr. Macleod AD et al., Mov Disord 2014. [source]
PD at 80: motor and non-motor complications accelerate. [source]
10yr Alzheimer's → late-stage. Xie J et al. BMJ 2008: median OS 4.5yr from dx. Temporal amplification (v3.8.0). [source]
5yr post-stroke: excess declines to HR ~1.3-1.5x. Hankey et al. Stroke 2000. [source]
5yr post-heart TX: conditional median survival ~10-15yr from dx. ISHLT Registry 2017. [source]
MS 15yr on modern DMTs: LE reduction ~2-5yr. Marrie RA, Lancet Neurol 2015; Koch-Henriksen 2010 (pre-DMT data, wider gap). [source]
7. Autoimmune, Hematologic, Infectious, GI
Newer conditions validated against registry and cohort study data.
Scale: 0–40 years remaining
ART-suppressed HIV: LE approaches general population (-2 to -5yr). Antiretroviral Therapy Cohort Collaboration, Lancet HIV 2017. [source]
Untreated/advanced AIDS CD4 <200: markedly reduced LE. CDC HIV surveillance reports. [source]
SLE SMR ~2-3x. Jorge AM et al., Arthritis Care Res 2018. Improved with modern immunosuppression. [source]
EUSTAR registry: 10-year survival ~65%. Elhai M et al., Ann Rheum Dis 2017. [source]
SCD median LE ~45-55yr (improving). Platt OS et al., NEJM 1994. Lanzkron S, Blood 2013. [source]
IPSS-R intermediate risk: median survival 3-5yr. Greenberg PL et al., Blood 2012. [source]
8. Comorbidity Combinations (Synergy Validation)
Tests condition-pair synergy amplification against published multi-morbidity outcomes.
Scale: 0–30 years remaining
DM2 + CHF: mortality 2-3x higher than CHF alone. MacDonald MR et al., Eur Heart J 2008. [source]
Triple comorbidity: multi-morbidity mortality data. Barnett K et al., Lancet 2012. [source]
SSc-PAH: 3-year survival ~50-60%. Condliffe R et al., AJRCCM 2009. [source]
COPD + CVD: ~30% increased mortality vs either alone. Chen W et al., Lancet Respir Med 2015. [source]
DM + CKD stage 4: 5-year mortality >50%. Afkarian M et al., JAMA 2013. [source]
Post-cardiac transplant CKD: major driver of late mortality. Ojo AO et al., NEJM 2003. [source]
Alcohol-related liver disease: SMR 3-5x. Rehm J et al., Lancet 2009. Synergistic mortality amplification. [source]
Depression + DM2: HR 1.5-2.0 vs DM2 alone. Katon WJ et al., Diabetes Care 2005. [source]
9. Clinical Scale Differentiation
Verifies that clinical grading scales (NYHA, GOLD, CKD) produce monotonically decreasing LE.
Scale: 0–30 years remaining
NYHA II: 5-year mortality ~20-30%. MAGGIC meta-analysis. [source]
NYHA III: 5-year mortality ~40-60%. [source]
CKD stage 2 (GFR 60-89): minimal mortality impact. Go AS et al., NEJM 2004. [source]
CKD stage 4: HR ~3.2 vs normal. Go AS et al., NEJM 2004. [source]
10. Typed Transplants
Organ-specific transplant outcomes validated against ISHLT, USRDS, and UNOS registry data.
Scale: 0–30 years remaining
ISHLT Registry: cardiac transplant median survival ~11yr. Lund LH et al., JHLT 2017. [source]
ISHLT Registry: older recipients have shorter post-transplant survival. [source]
ISHLT Registry: lung transplant median survival 6.7yr from transplant; moderate graft function at age 65 allows longer LE. Chambers DC et al., JHLT 2019. [source]
USRDS: 85% 5-year graft survival, near-normal LE for well-functioning grafts. Hart A et al., AJT 2021. [source]
UNOS: liver transplant 75% 5-year survival. Kim WR et al., Hepatology 2015. [source]
ISHLT: severe graft dysfunction/rejection markedly reduces post-transplant survival. [source]
ISHLT Registry: lung transplant survival decreases with recipient age. [source]
USRDS: older kidney transplant recipients have shorter but still improved LE vs dialysis. [source]
11. Behavioral Conditions
Depression and substance use disorder validated against meta-analyses and Lancet global burden studies.
Scale: 0–35 years remaining
Cuijpers P et al., 2014 meta-analysis: depression HR 1.52. LE reduction ~3-5yr. [source]
Severe depression: HR 2.0+. Walker ER et al., JAMA Psychiatry 2015. [source]
Degenhardt L et al., Lancet 2018: SUD SMR 2.0-4.0x. Drug/alcohol dependence LE loss 10-20yr. [source]
Severe SUD: SMR 3-5x. Degenhardt L et al., Lancet 2018. [source]
Dual diagnosis (SUD + MDD): synergistic mortality. Roerecke M & Rehm J, Addiction 2013. [source]
Depression at 70: HR ~1.5. Cuijpers P et al., 2014 meta-analysis. [source]
SUD at 65: SMR 2-4x. Degenhardt L et al., Lancet 2018. [source]
Severe SUD at 60: major LE reduction. Westman J et al., PLoS One 2015. [source]
12. New Conditions (v3.1)
Validates 8 new conditions added in engine v3.1: osteoporosis, valvular heart disease, anemia, epilepsy, osteoarthritis, vision impairment, prediabetes, family history cancer.
Scale: 0–35 years remaining
Abrahamsen B et al., Osteoporos Int 2015; T-score <-2.5 HR 1.4-1.7. [source]
Bliuc D et al., JAMA 2009; post-hip-fracture 1yr mortality 20-30%. [source]
Enriquez-Sarano M et al., NEJM 2005; moderate MR HR 2.0-2.5. [source]
Iung B et al., Eur Heart J 2003; severe pre-surgical valve disease. [source]
Penninx BW et al., JAMA 2004; elderly Hb<12 HR 1.6. [source]
Culleton BF et al., Blood 2006; Hb<10 HR 2.0-2.4 in elderly. [source]
Forsgren L et al., Epilepsia 2005; symptomatic epilepsy SMR 2.3-2.6. [source]
Neligan A et al., Epilepsia 2012; refractory SMR 4-6x, SUDEP 1/150/yr. [source]
Hawker GA et al., Arthritis Rheumatol 2014; adjusted HR 1.09-1.29. [source]
Nuesch E et al., BMJ 2011 meta-analysis; OA HR 1.55 (unadjusted). [source]
Knudtson MD et al., Arch Ophthalmol 2006 (Beaver Dam); VI HR 1.2-1.6. [source]
Lee DJ et al., Invest Ophthalmol Vis Sci 2002; VA <20/200 HR 1.6-1.9. [source]
Huang Y et al., BMJ 2016 meta-analysis; prediabetes HR 1.13 (95% CI 1.10-1.17). [source]
Barr EL et al., Diabetes Care 2007; IGT HR 1.2-1.3 in elderly. [source]
Mucci LA et al., Int J Cancer 2016; 2+ 1st-degree relatives, attenuated >60. [source]
Swedish Family-Cancer Database; BRCA/Lynch penetrance studies. [source]
13. Genetic / Family History
Parental longevity and early-death family history validated against twin/cohort heritability studies and parental-lifespan GWAS.
Scale: 0–40 years remaining
Ljungquist et al., J Gerontol 1998; Kaplanis et al., Science 2018; 20-30% heritability. [source]
Ljungquist et al., J Gerontol 1998; modest parental longevity signal. [source]
Pilling et al., Aging Cell 2017; parental lifespan GWAS. [source]
Sources & standards
We benchmark against the same standards your own actuaries and medical directors cite.
Reproducibility
Every estimate the platform produces is stamped with the engine version, a SHA-256 hash of the complete parameter set, and the Monte Carlo seed. Rerunning the same profile with the same version produces the same numbers — a property that matters for fiduciary documentation and audit. The snapshot on this page was generated from engine v3.9.0 on 2026-05-26.
Limitations
We are deliberate about what this engine is and is not:
- Published benchmarks are population-level; individual outcomes vary widely around any central estimate (which is exactly why we surface confidence intervals).
- Condition profiles are calibrated to published mortality ratios and survival data, not fitted to any single insurer's proprietary experience.
- Treatment compliance is assumed moderate unless the profile specifies otherwise.
- The engine is built for screening and planning, not for binding pricing. A formal life expectancy determination still requires medical-records review and, for life settlement, a certified LE provider. Lumis Life is not HIPAA-compliant and does not provide medical advice.
About the author
Lumis Life was built by Jeff Ting, a Fellow of the Society of Actuaries (FSA) and a CFA Charterholder who spent years pricing large-scale transactions at major insurance institutions. Read more on the about page, or see the underwriting-estimator methodology.