Research
Why healthspan extension was cut against the frontier base rate
Decomposed as a conjunction: roughly a 0.45 chance that a genuine ageing-slowing intervention is validated in humans early enough to show population effect
Researched Sep 10, 2026 · published Sep 13, 2026
What changed
Decomposed as a conjunction: roughly a 0.45 chance that a genuine ageing-slowing intervention is validated in humans early enough to show population effects by 2126, roughly 0.45 that it delivers twenty or more healthy years to typical people rather than to a high-risk subgroup, and roughly 0.7 that deployment and attribution are good enough for replicated population evidence. That product is near 0.14. The stated midpoint sits above it at 0.25 because a century is long enough for more than one attempt, because the decomposition assumes today’s mechanism classes are the only candidates, and because the resolution criterion still admits some interpretive latitude. The high end of 0.42 carries the case where partial reprogramming or a successor genuinely compresses morbidity; the low end of 0.12 carries the case where the deceleration is structural and geroscience never clears regulation.
Evidence that moved the estimate
- Rapamycin raised median mouse lifespan 23 to 26 percent across three independent sites; the same proportional effect on an eighty-year human life would be about eighteen years, so the mechanism class is not numerically absurd (hf-evidence-0007).
- Partial reprogramming acts on already-old animals rather than delaying onset, which is a qualitatively different lever from disease-by-disease medicine (hf-evidence-0010).
- Senolytics show the mechanism is druggable in humans, not only in animals (hf-evidence-0009).
- A modelled 1.9-year gain from semaglutide in a high-risk population shows a broad intervention can plausibly move population life expectancy at all (hf-evidence-0012).
- The horizon is a full century and the forecast needs only one success within it, so the current absence of translation is evidence about now rather than about 2126.
Evidence that held it back
- Frontier life expectancy improvement has decelerated below 0.2 years per year, and the authors of that analysis put the optimistic ceiling at 15 percent of females and 5 percent of males reaching 100 this century (hf-evidence-0013).
- The healthspan-lifespan gap widened from 8.5 to 9.6 years between 2000 and 2019, so added years are increasingly unhealthy ones (hf-evidence-0014).
- Global healthy life expectancy gained about 2.8 years per decade to 2019, largely through poorer countries converging on the frontier, and then gave back 1.6 years by 2021 (hf-evidence-0015).
- The slowdown appears simultaneously in all seventeen European countries studied, so it is structural rather than local (hf-evidence-0016).
- The flagship human trial of an ageing indication has not launched or enrolled and is still seeking funding, so the human evidence pipeline is stalled rather than merely immature (hf-evidence-0008).
- The best measured survival gain from a mass-deployed medicine is 12.6 days for statins, and the paper’s authors warn explicitly against extrapolating it to years (hf-evidence-0011).
- Translation, not mechanism discovery, is the acknowledged bottleneck, and no trial yet uses ageing as an endpoint (hf-evidence-0017).
What would move it next
A first adequately powered human trial that uses an ageing or composite healthspan endpoint reaching enrolment, and the first year in which healthy life expectancy at the frontier resumes rising faster than life expectancy.