Longevityresearch.ca Unveils a Unique Bayesian Causal Atlas; Saves up to 7.9 life years/patient
Delhi Scan/10337239

TORONTO - DelhiScan -- The Longevity Research Initiative today released results from a retrospective test of its causal-inference "oracles" against the U.S. Centers for Disease Control and Prevention's Third National Health and Nutrition Examination Survey (NHANES III). Across 15,774 adults — 3,500 of whom matched a modeled mortality-related condition — the analysis estimated a mean upper-bound gain of 7.9 life-years per patient relative to the care recorded in the dataset.

The Initiative was careful to characterize the figures as modeled ceilings rather than clinical predictions. "These numbers describe what a causal model says is achievable if every relevant intervention worked at the strongest effect size in the published literature, with the statistical double-counting between interventions removed," said Stuart Berkowitz, founder of the Longevity Research Initiative. "They are hypothesis-generating upper bounds, not promises to any individual patient — and we say so on every page."

Modeled results by disease domain

Within the 3,500 routed adults, the modeled upper-bound life-year gains were:

• Pulmonary disease (COPD, idiopathic pulmonary fibrosis, pulmonary arterial hypertension): +10.25 life-years
• Heart disease (cardiovascular mortality): +9.28 life-years
• Metabolic disease (type 2 diabetes): +8.88 life-years
• Cancer (cause-specific mortality): +4.54 life-years
• Brain (tumour and stroke): +3.65 life-years

For conditions without a mortality endpoint, the framework reports condition-specific outcomes instead of life-years. For osteoarthritis, drawn from the same NHANES adults, the model estimated up to an 82% reduction in pain at the maximum achievable intervention set.

More on Delhi Scan

• Kolbus Introduces the Next Step in Casemaking Efficiency
• Florida Law Advisers, P.A. Named Best Divorce Firm of 2026 by Expert Law Attorneys
• Sounds of LA County: 27 Parks.108 Concerts. One County
• Only One Flight Stands Between Los Angeles Youth Leaders and a Life-Saving Mission in South Africa
• Stigma Across Borders: Concerns Grow Over Discrimination Against Shincheonji Members Abroad

What makes the method distinctive

Everything is free, funded by the founder. Most risk calculators simply add or multiply individual risk reductions, which over-counts benefit when interventions act through the same biological pathway. The Initiative's tools instead implement the complete machinery of Judea Pearl's structural causal models: explicit removal of cross-correlation between interventions, E-value sensitivity analysis after VanderWeele and Ding, Tian-Pearl counterfactual probabilities of necessity and sufficiency, dose-response saturation curves, and Pareto "minimum-effective-set" optimization that identifies the smallest bundle of interventions capturing most of the achievable benefit.

The Initiative believes the resulting Bayesian Causal Atlas is the only public resource in the world today that assembles this full causal-inference toolkit across dozens of clinical and longevity domains and exposes it as free, interactive software — with each intervention adjustable by checkbox or dose slider and every effect size traceable to its source study.

Background: the Bayesian Causal Atlas

Hosted at longevityresearch.ca, the free Atlas comprises roughly 70 interactive analyses spanning cardiovascular disease, multiple cancers, kidney and liver disease, dementia and other brain pathologies, rare diseases, and all-cause mortality, among others. Each analysis pairs a browser-based dashboard with a formal report and is built on the same transparent engine, allowing users to construct an intervention bundle and watch the cross-correlation-corrected hazard ratio, projected life-years, and biological-pathway coverage update in real time. Ideal users are doctors and their patients.

The Initiative also released two new public tools: a healthy-longevity optimizer that ranks the highest-leverage interventions for extending lifespan, and an "optimal bundle" builder that runs the complete 45-intervention longevity model on the same causal engine.

More on Delhi Scan

• World Cup Crowds Are a Stress Test for America's Restrooms
• Postmortem Pathology Expands Access to Private Autopsy Services in Las Vegas
• How Sacramento Families Are Using Private Autopsies to Protect Inheritances, Resolve Insurance Claims, and Find Closure
• Los Angeles' Best Food: Food Journal Magazine Examines the Trends Shaping the City's Dining Scene
• Landmark Construction Expands Glass, Glazing, and Commercial Remodeling Services Across Los Angeles County and Surrounding Areas

Important limitations

The NHANES analysis is retrospective, modeled, and not peer-reviewed. Estimated life-year gains are deliberate upper bounds produced by applying best-case published hazard ratios to a general-population life table; real-world gains would be smaller and depend on individual baseline risk, adherence, and genetics. Cohort assignment relied on self-reported survey variables, an inherently imperfect routing method. Several interventions in the longevity tools — including senolytics, low-dose lithium, and antiviral strategies — are experimental or preclinical and are labeled as such. None of the Initiative's tools constitute medical advice, diagnosis, or treatment. Individuals should make decisions about medications, supplements, and lifestyle change only in consultation with a qualified clinician.

About the Longevity Research Initiative

The Longevity Research Initiative is a Toronto-based quantitative research effort that applies Judea Pearl's structural causal modeling to clinical and longevity questions, publishing its methods and tools openly at longevityresearch.ca. The Initiative's standard deliverable pairs an interactive dashboard with a formal technical report so that every assumption and effect size is inspectable.

Media Contact

Stuart Berkowitz, Founder, Longevity Research Initiative

Email: stuart@longevityresearch.ca Web: https://longevityresearch.ca