On U.S. life tables, eliminating all cancer adds only about 3 years to life expectancy at birth, and even eliminating all cardiovascular disease adds on the order of 3–7 years — because the survivors simply die of the next cause. The conditions modelled in this atlas account for roughly 91% of U.S. deaths (contributing-cause attribution), so optimising them together acts on nearly every cause at once. Applying the atlas’s mean per-patient modelled gain (+7.9 life-years) across that mortality burden, our life-expectancy-shift calculator projects a population life-expectancy increase of about +7.1 years — a modelled upper bound, anchored to the SSA 2022 life table. We do not know how to cure cancer or heart disease. We do know how to optimise the causes you can change — and that is available now.
A complete, source-verified catalogue of every cited reference in the atlas’s citation-bearing oracles — each line giving the reference, the oracle that uses it, its effect size and endpoint, and an A–D evidence rating — together with a full account of how the A–D rating system is constructed. Download here.
These analyses are evidence synthesis, not medical advice and not a prescription — the effect sizes describe trial populations, not you, and many interventions were studied in isolation rather than in combination. Use the atlas to inform a conversation with your physician, not to self-prescribe. Once you and your doctor settle on a candidate basket of interventions, check on the internet for contraindications and drug–drug interactions within the bundle and against any other drugs, supplements or conditions you already have — then change the bundle and repeat the check until you and your doctor are satisfied. No intervention here should be started, stopped or combined without that step.
Regulators including the FDA increasingly accept Bayesian methods as rigorous clinical evidence — the agency has issued guidance on the use of Bayesian statistics in clinical trials (notably medical-device trials) and in complex innovative trial designs. A randomized controlled trial (RCT) can cost millions of dollars and take years; a Bayesian causal synthesis of the trials already published runs for little more than the electricity to compute it. And because the number of possible intervention combinations is exponential — an intractable set for any physician to weigh by hand — we use computers to search it for the best set of interventions for each listed chronic illness, sub-type, and stage. On the cohorts above, the modelled gain is about 7.9 life-years for a person with a chronic illness. Please share this free website with the ~21% of your family and friends living with a chronic illness — and give the gift of life.
The analyses in this atlas address conditions that together account for 2.45 million annual deaths in the United States — 79.9% of all 2024 US deaths as catalogued by the CDC, including suicide, drugs, accidents and the 25 largest rare diseases (CDC: Leading Causes of Death). The figure is computed bottom-up from a per-analysis ICD-10 UCOD mapping with strict deduplication (each code owned by exactly one analysis); see the Atlas Coverage Count analysis for the full breakdown. Under contributing-cause attribution coverage rises to ~2.8 million (~91%). If Bayesian causal stacking of evidence-based interventions can compound their individual effects, the upper-bound life-years recovered are substantial.
The framework rests on the structural causal model of Judea Pearl, recipient of the Turing Award — computer science's equivalent of the Nobel Prize — for his work on probabilistic and causal reasoning. The same identification framework has been granted permission by the U.S. Food and Drug Administration for use in medical and drug-approval contexts, particularly for confounding adjustment in real-world evidence submissions and for handling counterfactual estimands in regulatory analyses.
The principal advantage over single-intervention thinking is the combined effect of a stack of interventions. Where a single agent may produce, say, a 28% reduction in disease burden, a properly de-correlated stack of mechanistically independent interventions can plausibly approach a 90% reduction on the same endpoint. This extends a clinician's toolkit by quantifying the joint effect of multi-modal protocols that no single randomised trial would test.
Any motivated reader can produce an equivalent analysis, in private, with greater detail tailored to their own condition, by sending the following prompt to a frontier large language model:
When doing a Bayesian causal analysis do the following: Using Judea Pearl's Bayesian Causal Analysis compute from all internet hazard ratios, mechanism of action and dose response studies for possible interventions to determine causal factors for all reductions in the stated end point. { insert your disease condition here } Make no assumptions about independence and remove all cross correlations. Do what-if, if-not-for, sensitivity, pareto and standard deviation/confidence level and all other analyses. Report in a PDF and an interactive HTML display where each intervention and choice is made with a check box or slider bar. Using dose response studies to determine saturation, discuss which % of cross correlations are appropriate for risk reduction.