The telos’s #1 modifiable exposure, finally quantified in the fabric. The wiki names smoking as the dominant big rock but held no source for it; Jha 2013 (216,917 US adults, NHIS linked to death records) supplies the contemporary magnitude — and, more decision-relevant, the shape of what cessation buys as a function of age. It is the exposure against which every nutrition lever in this vault is small.
The hazard: ~3x mortality, a decade of life
- Current smokers die at about three times the rate of never-smokers: all-cause HR 3.0 (99% CI 2.7-3.3) for women, 2.8 (2.4-3.1) for men, adjusted for age, education, alcohol, and adiposity. [EXTRACTED (Jha - Smoking Hazards Cessation Benefits 2013) chunk 01]
- At least a decade of life lost — «Life expectancy was shortened by more than 10 years» (~11 in women, ~12 in men). Survival to age 80 roughly halves: women 70% (never) vs 38% (current); men 61% vs 26%.
- ~60% of a smoker’s deaths (ages 25-79) are attributable to smoking. The excess is neoplastic + vascular + respiratory: lung cancer HR ~15-18, respiratory ~9, ischemic heart disease ~3.2-3.5, stroke up to 3.2.
This is the magnitude the Layer-1 ranking is built on: an HR of ~3 on all-cause mortality dwarfs the diet levers the vault otherwise holds (SFA replacement RR ~0.83 on CV events; a fibre or sugar shift smaller still). No attainable precision about meal timing changes what a smoker should do next -> Layer 1 - Ranking Interventions for a Stratum.
Cessation: steeply age-graded — the earlier, the more recovered
The benefit of quitting is large and depends heavily on when:
| Quit age | Years of life gained vs continuing | Residual all-cause HR vs never-smoker |
|---|---|---|
| 25-34 | ~10 | 1.0 — survival ~= never-smoker |
| 35-44 | ~9 | 1.2 |
| 45-54 | ~6 | 1.5 |
| 55-64 | ~4 | 1.7 |
| continue | — | 2.9 |
- Cessation before ~40 removes about 90% of the excess risk of death from continued smoking — «Cessation before the age of 40 years reduces the risk of death associated with continued smoking by about 90%.» [EXTRACTED (Jha - Smoking Hazards Cessation Benefits 2013) chunk 01]
- But it is not a clean reset — the “safe to smoke until 40” misread. Quitting by ~39 still left a ~20% excess risk (HR 1.2): «about one in six of these former smokers who dies before the age of 80 years would not have died» had they never smoked. So earlier is strictly better, and “quit by 40” is a floor on the benefit, not a licence to smoke until then.
- Even late cessation pays: quitting at 45-54 still removed roughly two-thirds of the excess risk. On a Layer-1 ranking there is no age at which quitting is not the largest available lever for a smoker.
Why the absolute benefit is large where the relative one is fixed
The authors make the baseline-risk argument directly: «Because the absolute risks of continuing to smoke are large, the absolute benefits of cessation will also be large» — and widening as never-smoker death rates fall. This is Baseline Risk and the Relative-Absolute Split in its cleanest form: a large relative effect (HR ~3) sitting on a high absolute baseline yields an enormous absolute gain, unlike the diet levers whose relative effects are small AND sit on lower baselines.
The sick-quitter correction — a worked reverse-causation fix
Jha handles the exact bias the The U-Shaped Association Artifact diagnostic warns about: «Life- threatening illness can cause smokers to quit, which distorts the rates of death among current smokers and among those who have quit smoking recently in opposite ways.» The fix — reclassify anyone who quit within 5 years of death as a current smoker, and check by excluding the first 2 years of follow-up (results unchanged) — is a clean template for separating a real protective effect from a frailty/sick-quitter artifact. It biases against overstating cessation’s benefit, so the large benefit survives the conservative handling.
Why confidence is high
- Effect size + dose + independence. An HR of ~3 is far too large to be plausibly confounded away;
adjustment for education/alcohol/adiposity «did little to alter the hazard ratios»; and the result is
concordant across four independent cohorts (British doctors, the Million Women Study, Thun’s 50-year
US series, a Japanese cohort). That triangulation — different populations, eras, and designs agreeing —
is why this sits at
confidence: highon a single ingested source: the source is one, but the backing it reports is not.
Limits
- Observational (no one randomizes smoking) — but the effect size, dose-response, adjustment stability, and cross-cohort concordance carry the causal read; this is the case where observational evidence is decisive.
- All-cause and cause-specific mortality; not trajectory/quality-of-life — the decade lost is a length claim; the shape of decline (COPD morbidity, post-stroke disability) is not quantified here.
- US population, cross-sectional smoking status at baseline — some baseline smokers later quit, which would understate the true hazard of continued smoking, not overstate it.
Self-critique [run 2026-07-29, before commit]
- Over-claim check: every magnitude is quoted or tabled from Jha; the “90% reduction” is paired with
its own residual (HR 1.2, ~20% excess) so it is not read as a clean reset;
confidence: highis justified by the source’s reported cross-cohort triangulation, not by a single study’s authority. - Reverse-causation: the sick-quitter correction is presented as biasing against the finding, so the surviving benefit is conservative — not laundered.
- Scope discipline: kept to mortality/life-expectancy (what Jha measures); did not import cessation methods (pharmacotherapy, behavioural support) — a prescriber act and a different question.